クラウド

2025.04.12

CSA STAR for AI 関係...

こんにちは、丸山満彦です。

Cloud Security Allianceが、AIのための監査プログラム(STAR for AI)を検討していますが、検討参加の締め切りが近づいているので...

検討に参加したい人は4月28日までに...

ちなみに、STARは Security, Trust, Assurance & Risk (セキュリティ、信頼、保証、リスク)の略ですね...、

 

考えることはだいたい似ています(^^)...

 

Cluod Security Alliance

An Urgent Need for AI Security Assurance

STAR Program for AI STAR Program for AI
Building AI trust upon STAR’s proven foundations STARの実証された基盤の上にAIの信頼を築く
An Urgent Need for AI Security Assurance AIセキュリティ保証の緊急ニーズ
“We all know we are playing with fire - everyone is trapped in this prisoner’s dilemma. Nobody is willing to stop the race to Superintelligence. Some believe that regulations are premature and potentially harmful. Others believe that in the absence of them, we’ll quickly lose control of society as we know it and go down the path of a dystopian future.”  「我々は皆、火遊びをしていることを知っている。誰もがこの囚人のジレンマに陥っている。誰も超知能への競争を止めようとはしない。規制は時期尚早であり、潜在的に有害であるという意見もある。ある者は、規制は時期尚早であり、有害である可能性があると考え、またある者は、規制がなければ、われわれが知っているような社会の制御をたちまち失い、ディストピア的な未来への道を歩むことになると考えている。
- Daniele Catteddu, CTO, CSA - CSA、CTO、Daniele Catteddu
We have firmly embedded generative AI technologies into our organizations and personal lives, and adoption is only increasing from there. However, at the same time, people face uncertainty and distrust with AI. No common standard exists. Various groups are slowly building AI frameworks, but their pace clashes with the light-speed progress of AI. 私たちは生成的AI技術を組織や個人生活にしっかりと組み込んでおり、そこから採用は増える一方だ。しかし同時に、人々はAIに対する不確実性と不信感に直面している。共通の標準は存在しない。様々なグループがAIの枠組みを少しずつ構築しているが、そのペースはAIの光速の進歩と衝突している。
Launched in 2011, CSA’s Security, Trust, Assurance & Risk (STAR) program is the industry’s most powerful program for security assurance, listing over 3,400 cloud provider security assessments in its publicly-available registry. Worldwide, organizations rely on STAR entries as an indicator of cloud service trustworthiness. This framework is the perfect medium for bringing stability to the chaotic realm of AI. 2011年に開始されたCSAのセキュリティ、トラスト、アセスメント&リスク(STAR)プログラムは、セキュリティ保証のための業界で最も強力なプログラムであり、一般公開されているレジストリに3400以上のクラウドプロバイダのセキュリティ評価が登録されている。世界中の企業が、クラウドサービスの信頼性を示す指標としてSTARのエントリーを信頼している。この枠組みは、AIという混沌とした領域に安定性をもたらすのに最適な媒体である。
CSA is expanding the STAR program to include assurance for AI. STAR for AI will pull from the Cloud Controls Matrix (CCM) and other existing auditing standards to deliver a security framework and certification program for AI services, as soon as possible. This will provide AI companies, cloud providers, and enterprise users with an authoritative mechanism to measure AI trustworthiness. CSAはSTARプログラムを拡張し、AIの保証を含める。STAR for AIは、Cloud Controls Matrix(CCM)やその他の既存の監査標準を活用し、AIサービス向けのセキュリティフレームワークと認証プログラムを早急に提供する。これにより、AI企業、クラウドプロバイダー、エンタープライズユーザーに、AIの信頼性を測定する権威あるメカニズムを提供する。
What Are We Building? 何を構築するのか?
AI Controls Matrix AI Controls Matrix
A framework of control objectives to support the secure and responsible development, management, and use of AI technologies. Draws from the Cloud Controls Matrix (CCM), ISO/IEC 42001, ISO 27001, and more.  AI技術の安全で責任ある開発、管理、利用をサポートするための管理目標の枠組み。クラウドコントロールマトリックス(CCM)、ISO/IEC 42001、ISO 27001などから引用している。
AI Safety Pledge AI Safety Pledge
A list of high-level AI safety principles for companies to pledge to support. Serves as a stepping stone to the broader certification program and draws from the AI Controls Matrix. 企業が支持を表明するためのハイレベルなAI安全原則のリスト。より広範な認証プログラムへの足がかりとなるもので、AI Controls Matrixを活用している。
AI Auditing Scheme AI Auditing Scheme
Conformity assessment mechanisms to evaluate the adherence of AI services to the AI Controls Matrix. Possible approaches include self-assessment, third-party audit, continuous controls monitoring, and AI auditing. AIサービスのAIコントロール・マトリックスへの準拠を評価するための適合性評価メカニズム。自己アセスメント、サードパーティ監査、継続的管理モニタリング、AI監査などのアプローチが考えられる。
AI Safety Certification Program AI Safety Certification Program
A certification service delivered via the cloud that will leverage the AI Controls Matrix and live on the CSA STAR Registry. AI Controls Matrixを活用し、CSA STAR Registry上で稼働するクラウド経由で提供される認証サービス。
Stay Informed 情報提供
People around the world are calling for the prompt regulation of AI services. CSA has taken up the mantle to deliver.  世界中の人々がAIサービスの迅速な規制を求めている。CSAはそれを実現するためにマントルを引き受けた。
Fill out this form to stay updated as we develop STAR for AI. Be the first to hear about new calls for participation, peer reviews, and releases. After submitting, you’ll get access to additional resources to explore the initiative further. このフォームに入力すると、AI向けSTARの開発に関する最新情報を入手できる。新しい参加募集、ピアレビュー、リリースに関する情報をいち早くお届けする。送信すると、このイニシアチブをさらに探求するための追加リソースにアクセスできるようになる。

 

ブログ...

・2025.01.29 Can GenAI Services Be Trusted? | At the Discovery of STAR for AI

 

Can GenAI Services Be Trusted? | At the Discovery of STAR for AI GenAIサービスは信頼できるか?| GenAIのサービスは信頼できるのか|AIのためのSTARの発見
Written by Daniele Catteddu, Chief Technology Officer, CSA. CSA最高技術責任者、Daniele Cattedduが書いた。
Whenever new technologies are introduced into our personal lives, organizations, or even society as a whole, we always ask the same question: Can I trust it? 新しいテクノロジーが私たちの個人生活や組織、あるいは社会全体に導入されるとき、私たちはいつも同じ質問をする: それは信頼できるのか?
Most recently, how many of us have asked ourselves whether, how, and within which limits we can trust LLM/GenAI services and their providers? This is a legitimate question, given that we are faced with an innovation that holds the potential to transform our lives profoundly. Policymakers and regulators seem to be grappling with similar concerns. Meanwhile, GenAI service providers are asking: How can we earn the trust of our customers, policymakers, regulators, and markets? 最近では、LLM/GenAIのサービスやそのプロバイダを信頼できるのか、どのように、どの範囲内で信頼できるのか、どれだけの人が自問しているだろうか。私たちの生活を大きく変える可能性を秘めたイノベーションに直面していることを考えれば、これは正当な疑問である。政策立案者や規制当局も同様の懸念に取り組んでいるようだ。一方、GenAIサービス・プロバイダは、「どうすれば顧客、政策立案者、規制当局、市場のトラストを獲得できるのか?
The same question was posed during the early days of cloud computing: Can I trust the cloud? Back then, the Cloud Security Alliance (CSA) brought together the expertise of providers, customers, auditors, and regulators to address the complex matters of trust and assurance in cloud computing, and we created the STAR Program. Building on that experience, CSA has introduced STAR for AI—a pioneering initiative inspired by those early days of cloud computing. クラウド・コンピューティングの黎明期にも同じ問いが投げかけられた: クラウドは信頼できるのか?当時、クラウド・セキュリティ・アライアンス(CSA)は、プロバイダ、顧客、監査人、規制当局の専門知識を結集して、クラウド・コンピューティングにおける信頼と保証の複雑な問題に取り組み、STARプログラムを創設した。その経験に基づき、CSAは、クラウド・コンピューティングの黎明期に着想を得た先駆的な取り組みであるSTAR for AIを導入した。
I’m Daniele Catteddu, CTO at CSA and Co-Founder of the STAR Program. In this blog, I will discuss current GenAI services governance, trust, risk management, and compliance challenges and will introduce the STAR for AI initiative that aims to establish a standard for GenAI services assurance. 私はCSAのCTOであり、STARプログラムの共同創設者であるDaniele Cattedduだ。このブログでは、現在のGenAIサービスのガバナンス、トラスト、リスクマネジメント、コンプライアンスの課題について説明し、GenAIサービス保証の標準確立を目指すSTAR for AIイニシアチブについて紹介する。
What does it mean for a GenAI service provider to be trustworthy? GenAIサービスプロバイダが信頼に足るとはどういうことか?
Or even better, what does it mean for a GenAI service to be trustworthy? あるいはそれ以上に、GenAIサービスが信頼に足るとはどういうことか?
For our purpose, we’ll define a Trustworthy GenAI Service as one that is committed to serving humanity responsibly and ethically via safe technologies. 我々の目的のために、信頼できるGenAIサービスとは、安全なテクノロジーを通じて、責任を持って倫理的に人類に奉仕することにコミットしているものと定義する。
It is offered by a system that is robust, reliable, resilient, explainable, controllable, transparent, accountable, protects privacy, fosters fairness, and complies with all applicable laws and regulations. それは、堅牢で、信頼性があり、レジリエンシーがあり、説明可能で、制御可能で、透明性があり、説明責任を果たし、プライバシーを保護し、公平性を育み、適用法および規制を遵守するシステムによって提供される。
Back in the early days of cloud computing, the challenges were: クラウド・コンピューティングの黎明期には、次のような課題があった:
・The introduction of a new business and technological paradigm (someone else’s computer syndrome) ・新しいビジネスと技術のパラダイムの序文(他人のコンピューター症候群)
・Confusion on how to govern data (where is my data in the cloud obsession) ・データをどのようにガバナンスするかについての混乱(クラウドの中の私のデータはどこにあるのかという強迫観念)
・Confusion on how to govern new categories of services (SaaS Governance nightmare) ・新しいカテゴリーのサービスをどのようにガバナンスするかについての混乱(SaaSガバナンスの悪夢)
・Confusion about how to use and integrate the new technology and services into existing platforms and whether to revolutionize existing habits, systems, and platforms based on the new technology ・新しいテクノロジーとサービスをどのように利用し、既存のプラットフォームに統合するか、また既存の習慣を変革するかどうかについての混乱、 
・Confusion in the legal and regulatory landscape ・法規制の混乱
・Lack of standards (both technical and assurance/quality standards) and need to retrofit existing ones ・標準(技術標準と保証・品質標準の両方)の欠如と既存標準の改修の必要性
For AI, the situation appears to be similar in many ways, with a few notable differences. The paradigm shift does not only affect businesses and technologies; it goes deeper into the very fabric of our society and personal lives. The new syndrome is not someone else’s computer but someone else’s (else’s) brain. The risk of over-reliance on the AI Oracle did not exist with the cloud. AIについては、いくつかの顕著な違いはあるものの、状況は多くの点で類似しているようだ。パラダイムシフトは、ビジネスやテクノロジーに影響を与えるだけでなく、私たちの社会や個人生活そのものに深く入り込んでいる。新しい症候群は、他人のコンピューターではなく、他人の(他人の)脳である。AIオラクルに過度に依存するリスクは、クラウドには存在しなかった。
We have the same confusion regarding data governance (the obsession with “who is using my data for training?”) and service governance (How to prevent errors, abuses, and unwanted uses? How to understand what is real and what is fake? What is human vs. AI-generated? Shadow-AI paranoia, Evil-GTP fear, soon-to-be Evil AI Agent fear). データガバナンス(「誰が私のデータをトレーニングに使っているのか」というこだわり)やサービスガバナンス(エラーや乱用、望まない利用をどう防ぐか)に関しても、同じような混乱がある。何が本物で何が偽物かを理解するにはどうすればいいか?人間対AI生成的とは何か?シャドーAIパラノイア、Evil-GTPの恐怖、もうすぐ登場するEvil AI Agentの恐怖)。
Similarly, many organizations face uncertainty, ranging between full embracement, timid acceptance, refusal, and complete denial. 同様に、多くの組織は、完全な受け入れ、臆病な受け入れ、拒否、完全な否定の間の不確実性に直面している。
Should I stay or should I go? The same old-school human resistance to change and legacy problem, exacerbated by the much stronger disruptive potential of GenAI compared to the cloud and mitigated by the substantially more advanced capabilities offered by GenAI. 残るべきか、去るべきか?変化やレガシー問題に対する昔ながらの人間の抵抗は、クラウドと比較してGenAIの破壊的潜在力がはるかに強いために悪化し、GenAIが提供する実質的により高度な機能によって緩和される。
What’s the current state of AI security assurance and trust? AIのセキュリティ保証と信頼の現状は?
The political, legal and regulatory landscape is tough. The geo-strategic interests at stake are at a scale humanity has seen only during the nuclear race during WW2. Acknowledging GenAI as a potential existential threat underscores the urgency of robust governance. However, I recognize that framing it in terms of the nuclear race may hinder productive dialogue by evoking a zero-sum, adversarial mindset. Instead, GenAI's risks should be addressed through a combination of cooperative international agreements, transparent development practices, and enforceable safety standards. 政治、法律、規制の状況は厳しい。地政学的な利害が絡んでおり、人類が第2次世界大戦中の核開発競争でしか見たことのない規模だ。GenAIを潜在的な脅威と認識することは、強固なガバナンスの緊急性を強調する。しかし、核開発競争という観点からこの問題をとらえることは、ゼロサム的で敵対的な考え方を想起させ、生産的な対話を妨げる可能性があることを私は認識している。その代わりに、GENAIのリスクは、協力的な国際協定、透明性のある開発慣行、強制力のある安全標準の組み合わせを通じて対処されるべきである。
We all know we are playing with fire - everyone is trapped in this prisoner’s dilemma. Nobody is willing to stop the race to Superintelligence. Some believe that regulations are premature and potentially harmful. Others believe that in the absence of them, we’ll quickly lose control of society as we know it and go down the path of a dystopian future. 囚人のジレンマに陥っているのだ。誰もがこの囚人のジレンマに陥っているのだ。誰も超知能への競争を止めようとはしない。規制は時期尚早であり、潜在的に有害だと考える者もいる。また、規制がなければ、私たちが知っているような社会のコントロールはたちまち失われ、ディストピア的な未来への道を歩むことになると考える人もいる。
The EU, China, and Brazil have already adopted regulations: the EU AI Act, Generative AI Regulations (e.g. Measures for the Administration of Generative Artificial Intelligence Services) and Brazilian AI Bill, respectively. The EU AI Act, which I know a bit better, aims to strike a good balance between innovation, safety, and accountability. At the same time, the USA government appears to espouse a self-regulatory approach. In between, there are several other countries in the process of figuring out the principles and laws to rule and guide the transition toward a new AI-driven society. EU、中国、ブラジルはすでに、それぞれEU AI法、生成的AI規則(生成的人工知能サービス管理措置など)、ブラジルAI法案という規制を採用している。EUのAI法は、イノベーション、安全性、説明責任のバランスを取ることを目的としている。同時に、アメリカ政府は自主規制的なアプローチを支持しているようだ。その中間には、新しいAI駆動型社会への移行をルール化し導くための原則や法律を見つけ出そうとしている国がいくつかある。
In the background, there is also a battle between parts of the scientific community. One side has the urge to slow down, or even pause, the exponential progress of AI until we have more solid solutions to the alignment problem. Others firmly believe we need to immediately commercialize these new technologies to build an AI-aware society and favor the embracement of the positive effect of GenAI. その背景には、科学界の一部による争いもある。一方は、アライメント問題に対するより確かな解決策が得られるまで、AIの指数関数的な進歩を遅らせたい、あるいは一時停止させたいという衝動に駆られている。他方では、AIを意識した社会を構築するためには、これらの新技術を直ちに商業化する必要があると固く信じており、GenAIのポジティブな効果を受け入れることを支持している。
The standards are clearly in their infancy. We are slowly building the vocabularies and taxonomies to understand each other. However, we can anticipate that the consensus-building exercise that stands behind the standard creation seems to clash with the light-speed progress of AI technologies and the continuous release of solutions. 標準は明らかに初期段階にある。互いを理解するための語彙や分類法が徐々に構築されつつある。しかし、標準作成の背後にある合意形成の運動は、AI技術の光速の進歩や継続的なソリューションのリリースと衝突するように思えることが予想される。
Someone could summarize the situation as a bit chaotic... この状況を、ちょっと混沌としている......と表現する人がいるかもしれない。
What is CSA going to do about it? CSAはどうするのか?
In this chaos, at CSA, we decided to use the same principles, tips, and tricks that proved useful for cloud computing assurance to build an approach that can offer a solid, robust, understandable, and measurable “something” —a starting point— to reiterate and improve continuously. Over time, CSA hopes to achieve a better approximation of the measure of trustworthiness of the AI services that we’ll consume. この混沌の中で、CSAでは、クラウド・コンピューティングの保証に有用であることが証明されたのと同じ原則、ヒント、トリックを使用して、反復し、継続的に改善するための、堅実で、強固で、理解可能で、測定可能な「何か」(出発点)を提供できるアプローチを構築することにした。時間をかけて、CSAは、我々が利用するAIサービスの信頼性の尺度のより良い近似値を達成したいと考えている。
This little “something” we are building is called STAR for AI. For those not familiar with CSA, STAR is the acronym for Security, Trust, Assurance, and Risk. STAR is CSA’s program for cybersecurity assurance, governance, and compliance, and its initial focus and scope was cloud computing. 我々が構築しているこの小さな「何か」は、STAR for AIと呼ばれている。CSAをよく知らない人のために説明すると、STARはSecurity(セキュリティ)、Trust(信頼)、Assurance(保証)、Risk(リスク)の頭文字をとったものだ。STARはサイバーセキュリティの保証、ガバナンス、コンプライアンスのためのCSAのプログラムで、当初はクラウド・コンピューティングに焦点を当てていた。
In the cloud computing industry, the STAR program is widely used as an indicator of cloud service trustworthiness, both in the private and public sector domains. For instance, some countries like Italy officially use it as a mechanism of adherence to national requirements for the public sector and critical infrastructures. クラウド・コンピューティング業界では、STARプログラムはクラウド・サービスの信頼性を示す指標として、民間・公共セクターを問わず広く利用されている。例えば、イタリアのように、公的部門や重要インフラに対する国家要件の遵守を示すメカニズムとして公式に使用している国もある。
And we hope to achieve the same success with our latest initiative. そして我々は、最新のイニシアチブでも同様の成功を収めたいと考えている。
STAR for AI will focus on Multimodal GenAI services and consider their overall value chain. This means the evaluation will have within its scope one or more of the following components: STAR for AIは、マルチモーダルなGenAIサービスに焦点を当て、そのバリューチェーン全体を検討する。これは、評価の範囲に以下の構成要素の1つ以上が含まれることを意味する:
1) Cloud / Processing Infra / GenAI Operations, 2) Models, 3) Orchestrated Services, 4) Applications, and 5) Data (for reference please read the CSA LLM Threat Taxonomy). 1) クラウド/処理インフラ/GenAIオペレーション、2) モデル、3) オーケストレーテッド・サービス、4) アプリケーション、5) データ(参考のため、CSA LLM脅威分類法をお読みください)。
Additionally, the program will evaluate the security of AI service usage. さらに、このプログラムでは、AIサービス利用のセキュリティを評価する。
Our goal is to create a trustworthy framework for key stakeholders to demonstrate safety and security, whether they are a Gen-AI frontier model owner, (added-value) AI service developers and providers, or AI services business users. 我々の目標は、主要なステークホルダーが、Gen-AIフロンティアモデルオーナー、(付加価値のある)AIサービス開発者やプロバイダ、AIサービスビジネスユーザーのいずれであっても、安全性とセキュリティを実証するための信頼できる枠組みを構築することである。
The program will focus on technical and governance aspects related to cybersecurity. Additionally, it will cover aspects of safety, privacy, transparency, accountability, and explainability as far as they relate to cybersecurity. このプログラムは、サイバーセキュリティに関連する技術的側面とガバナンスの側面に焦点を当てる。さらに、サイバーセキュリティに関連する限りにおいて、安全性、プライバシー、透明性、説明責任、説明可能性の側面もカバーする。
The audit and evaluation approach will be risk-based, i.e., the suitability of the controls in place will be established based on the risks to which the service is exposed. The risk-based approach will ensure the audit and evaluation process is relevant to the context and use case. 監査と評価のアプローチは、リスクベース、すなわち、サービスがさらされているリスクに基づいて、実施されているコントロールの適切性が確立される。リスクベースのアプローチは、監査・評価プロセスがコンテキストとユースケースに関連していることを保証する。
The program will leverage existing auditing and accreditation standards, which might be AI-specific (e.g., ISO/IEC 42001-2023) or more general (ISO27001, ISO17021, 17065, etc.). It will include both point-in-time and continuous auditing. このプログラムは、AIに特化した標準(ISO/IEC 42001-2023など)またはより一般的な標準(ISO27001、ISO17021、17065など)の既存の監査・認定標準を活用する。これには、ポイント・イン・タイム監査と継続的監査の両方が含まれる。
Introducing the AI Controls Matrix AIコントロール・マトリックスの導入
As a first step, we are establishing a control framework for GenAI service security. The framework is open, expert-driven, consensus-based, and vendor-agnostic. Its ambition is to repeat and improve on the success of the Cloud Controls Matrix and become an industry de facto standard. 第一段階として、GenAIサービス・セキュリティのためのコントロール・フレームワークを確立する。この枠組みは、オープンで、専門家主導で、コンセンサスに基づき、ベンダーにとらわれない。このフレームワークは、クラウドコントロールマトリックスの成功を繰り返し、改善し、業界のデファクトスタンダードとなることを目指している。
The framework is called the CSA AI Controls Matrix (AICM). この枠組みはCSA AI Controls Matrix(AICM)と呼ばれている。
We released it for peer review just before the end of 2024, hoping to offer our community a present for the beginning of the new year. 我々は、2024年末の直前にピアレビューのためにこのフレームワークをリリースした。
The initial version of AICM was designed to follow common sense principles of trustworthiness. We defined a taxonomy, created threat scenarios, and identified control objectives to mitigate the threats. The AI control objectives were framed in the CSA Cloud Control Matrix’s template and structure. We leveraged the strengths of the CCM model, and customized and improved where necessary. The team of volunteers contributing includes experts from industry, academia, and governmental bodies. Needless to say we also used the support of GenAI tools. AICMの初期バージョンは、信頼性の常識的な原則に従って設計された。我々は分類法を定義し、脅威シナリオを作成し、脅威を緩和するための制御目標を特定した。AIの管理目標は、CSAクラウド・コントロール・マトリックスのテンプレートと構造で組み立てられた。我々はCCMモデルの長所を活用し、必要に応じてカスタマイズと改善を行った。貢献したボランティアチームには、産業界、学界、政府団体の専門家が含まれている。GenAIツールのサポートも活用したことは言うまでもない。
The current version of the AICM is structured in 18 domains. 17 are in common with the CCM, plus the addition of the Model Security Domain. There are 242 control objectives (37 of them AI-specific, 183 relevant to both AI and cloud, and 22 cloud-specific). AICMの現在のバージョンは、18のドメインで構成されている。17はCCMと共通で、加えてモデル・セキュリティ・ドメインが追加されている。242の管理目標がある(そのうち37はAI固有、183はAIとクラウドの両方に関連、22はクラウド固有)。
This is our first call for action. If you care about AI service trustworthiness and/or are a subject expert on the matter, you should review the current draft of the AICM and contribute to improving it. It might seem like a hyperbole, but having your name on the standard that will design the boundaries of GenAI service trustworthiness, means leaving a legacy in history. これは、私たちが最初に呼びかける行動である。もしあなたがAIサービスの信頼性に関心があり、かつ/またはこの問題の専門家であるならば、AICMの現在のドラフトを見直し、改善に貢献すべきである。大げさに思えるかもしれないが、GenAIサービスの信頼性の境界を設計する標準にあなたの名前を載せることは、歴史に遺産を残すことを意味する。
Determining assessment mechanisms アセスメントメカニズムの決定
The other foundational component of STAR for AI is the auditing scheme, the conformity assessment mechanisms that will be used to evaluate the adherence of a given GenAI service to the AICM requirements. STAR for AIのもう一つの基礎となるコンポーネントは、監査スキームであり、与えられたGenAIサービスがAICM要件に準拠しているかどうかを評価するために使用される適合性評価メカニズムである。
With the support of the auditing and assurance community, in 2024 we started to reason about the mechanisms fit for GenAI service cyber assurance evaluation. More in general, we started a discussion on the impact of AI on auditing and the future of assurance and trustworthiness. 監査・保証コミュニティの支援により、2024年に我々はGenAIサービスのサイバー保証評価に適したメカニズムについて推論を開始した。より一般的には、AIが監査に与える影響、保証と信頼性の将来についての議論を開始した。
We are exploring options. Several possible approaches are considered for use. Some are already existing and standardized (self-assessment, third-party audit, etc.), others are under development (continuous controls monitoring/auditing), and others might be introduced as a result of new technologies or new technical needs (e.g., Gen AI Auditing/Assessment). 我々は選択肢を模索している。いくつかの可能性のあるアプローチを使用することを検討している。すでに存在し標準化されているもの(自己評価、サードパーティ監査など)、開発中のもの(継続的な統制モニタリング/監査)、新技術や新たな技術的ニーズの結果として導入される可能性のあるもの(Gen AI Auditing/Assessmentなど)などがある。
Here comes our second call for action, and once again, it involves you taking a step forward, being a thought leader, and contributing to shaping the future of cyber assurance. If you would like to be involved in the creation of the auditing scheme, please get in touch. また、新技術や新たな技術的ニーズの結果として導入される可能性のあるものもある(例:Gen AI監査/評価)。ここでもまた、あなたが一歩を踏み出し、オピニオンリーダーとして、サイバーアシュアランスの未来を形作ることに貢献することが求められている。監査スキームの構築に関わりたい方は、ぜひご連絡いただきたい。
An urgent call to action 緊急の呼びかけ
Why are we building STAR for AI? Simple: Within our community, there’s a background voice that is increasing in volume by the second. It quickly became a scream. What we hear is a request to support the controlled adoption of GenAI services and the governance of its cybersecurity and safety. We need a mechanism to measure trustworthiness, and CSA is strategically committed to delivering such a solution. なぜ我々はAIのためのSTARを構築しているのか?単純なことだ: 我々のコミュニティーの中で、刻一刻とそのボリュームを増している背景の声がある。それはすぐに悲鳴となった。私たちが耳にするのは、GenAIサービスの制御された導入と、そのサイバーセキュリティと安全性のガバナンスをサポートしてほしいという要望だ。信頼性を測定するメカニズムが必要であり、CSAはそのようなソリューションを提供することに戦略的にコミットしている。
To conclude, there is growing urgency in the market to establish reliable assurance and compliance mechanisms for governing GenAI services. This challenge is particularly complex as it intersects with broader ethical considerations and complex technology. 結論として、GenAIサービスをガバナンスするための信頼できる保証とコンプライアンスのメカニズムを確立することが市場で急務となっている。この課題は、より広範な倫理的考察や複雑なテクノロジーと交差するため、特に複雑である。
We face a notable paradox: even as we work to define parameters and metrics for GenAI trustworthiness, these technologies are already embedded in our organizations and personal lives. Adoption is only accelerating as organizations recognize the opportunities GenAI creates. Moreover, we are increasingly relying on GenAI tools for assessment and auditing processes, including autonomous decision-making. This creates the potential situation where we might depend on a technology to evaluate its own trustworthiness before we have established reliable methods to measure the integrity of the decisions the technology may take without human intervention. 我々がGenAIの信頼性のパラメータと測定基準を定義しようと努力している間にも、これらのテクノロジーは既に我々の組織や個人生活に組み込まれている。GenAIが生み出す機会を組織が認識するにつれ、採用は加速する一方だ。さらに、自律的な意思決定を含むアセスメントや監査プロセスにおいて、GenAIツールへの依存度が高まっている。このことは、人間の介入なしにテクノロジーが行う意思決定の完全性を測定する信頼できる方法が確立される前に、テクノロジー自身の信頼性を評価するためにテクノロジーに依存する可能性があるという状況を生み出す。
While this situation doesn't call for panic, it does demand urgent attention. I encourage all of you to contribute your expertise to the STAR for AI initiative to help address these critical challenges. この状況はパニックを引き起こすものではないが、緊急の注意が必要である。このような重大な課題に対処するため、STAR for AIイニシアティブに専門知識を提供していただきたい。

 

・2025.02.17 AI Controls Matrix

公開期間がおわったので、今はみられません...

 

20250412-63724

 

 

 

| | Comments (0)

2025.04.09

米国 一般調達局 FedRAMP 20X (2025.03.24)

こんにちは、丸山満彦です。

2024年にFedRAMPのプロセスが改訂され、FedRAMP 20Xに変わっていきますよね...

FedRAMP認可には費用も時間もかかりすぎる...ということでの改革ですね。 日本もFedRAMPを参考にISMAPを作っているので、同じ方向で改善していくのがよいのでしょうね...

ポイントは評価の自動化です...

 

U.S. General Services Administration

1_20250409061101

 

・2025.03.24 GSA announces FedRAMP 20x

U.S. General Services Administration 米国一般調達局
GSA announces FedRAMP 20x 一般調達局がFedRAMP 20xを発表
Implementing a new approach to accelerate cloud adoption クラウド導入を加速させる新たなアプローチを導入
WASHINGTON — The U.S. General Services Administration (GSA), a leader in providing tools and guidance to help federal agencies deliver seamless digital services to American taxpayers, announced today that the Federal Risk and Authorization Management Program (FedRAMPⓇ) will focus on working with industry to develop a new, cloud-native approach to authorizations. ワシントン - 連邦政府機関が米国の納税者にシームレスなデジタルサービスを提供するためのツールやガイダンスを提供するリーダーである米国一般調達局(GSA)は本日、連邦リスク・認可マネジメント・プログラム(FedRAMPⓇ)が業界と協力して認可に対する新たなクラウドネイティブ・アプローチの開発に注力することを発表した。
FedRAMP 20x will focus on innovating alternative approaches to make automated authorization simpler, easier, and cheaper while continuously improving security. The FedRAMP team will also continue to support traditional agency authorizations. FedRAMP 20xは、セキュリティを継続的に改善しながら、自動認可をよりシンプル、簡単、安価にするための代替アプローチの革新に重点を置く。FedRAMPチームは、従来の認可機関も引き続きサポートする。
“Since the start of this Administration, we’ve focused on improving government operations to make them more efficient and effective for every employee and the American taxpayer,” said GSA Acting Administrator Stephen Ehikian. “Our partnership with the commercial cloud industry needs serious improvement. Strengthening this relationship will help us fulfill our commitment to cutting waste and adopting the best available technologies to modernize the government’s aging IT infrastructure. FedRAMP 20x will give agencies access to the latest technology now — not months or years down the road.” 「ガバナンスの開始以来、我々は政府業務の改善に重点を置き、すべての職員と納税者にとってより効率的で効果的な政府業務を目指してきた。「商用クラウド業界とのパートナーシップは深刻な改善が必要だ。この関係を強化することは、無駄を削減し、政府の老朽化したITインフラを近代化するために利用可能な最善の技術を採用するという我々のコミットメントを果たすことにつながる。FedRAMP 20xは、数カ月後や数年後ではなく、今すぐに最新技術にアクセスできるようにする。
FedRAMP 20x is built on these core principles: FedRAMP 20x はこれらの基本原則の上に構築されている:
GSA will set the foundation for private sector innovation: FedRAMP will make it easier for cloud providers to follow modern security practices and show their leadership in developing secure cloud solutions for the government. FedRAMP will also hold public working groups to gather input from industry, ensure equal access to information, encourage pilot programs, and provide technical guidance before formal public comment and release. GSAは民間部門の技術革新の基礎を築く: FedRAMPは、クラウド・プロバイダが最新のセキュリティ慣行に従うことを容易にし、政府向けの安全なクラウド・ソリューションの開発においてリーダーシップを発揮できるようにする。FedRAMPはまた、業界からのインプットを集め、情報への平等なアクセスを確保し、パイロット・プログラムを奨励し、正式なパブリック・コメントとリリースの前に技術ガイダンスを提供するために、公開ワーキング・グループを開催する。
Cutting red tape through automation: FedRAMP is required for all federal agency cloud services, but getting approved currently involves a lot of paperwork and a slow manual process. FedRAMP 20x will reduce unnecessary paperwork and aim to automate as much of the process as possible to accelerate approvals in a cost efficient manner. 自動化によってお役所仕事を削減する: FedRAMPはすべての連邦政府機関のクラウド・サービスに義務付けられているが、現在、承認を得るには多くの書類作成と時間のかかる手作業が必要である。FedRAMP 20xでは、不必要なペーパーワークを削減し、可能な限りプロセスを自動化して、コスト効率の高い方法で承認を加速することを目指す。
Faster, more secure cloud adoption: Currently, it can take months or even years for a cloud provider to get FedRAMP approval, which slows down how quickly agencies can adopt new technology. FedRAMP 20x will simplify and clarify security requirements so that new cloud services can be approved in weeks instead of years. より迅速で安全なクラウド導入: 現在、クラウドプロバイダーがFedRAMPの承認を得るには数カ月から数年かかることもあり、そのため各省庁が新技術を採用するスピードが遅くなっている。FedRAMP 20x では、セキュリティ要件が簡素化・明確化されるため、新しいクラウド・サービスを数年ではなく数週間で承認できるようになる。
More flexibility and better collaboration: FedRAMP has traditionally acted as a middleman between agencies and cloud providers. FedRAMP 20x will build on existing trust and make it easier for providers and agencies to work together directly. より柔軟でより良いコラボレーション: FedRAMPは従来、省庁とクラウドプロバイダの仲介役として機能してきた。FedRAMP 20x は既存の信頼関係を基礎とし、プロバイダと各省庁が直接協力しやすくする。
“FedRAMP is a shared service that meets the critical needs of agencies government-wide,” said Technology Transformation Services Director and Deputy Commissioner of the Federal Acquisition Service Thomas Shedd. “We’re not just modernizing a process; we’re reimagining how federal cloud security can work and providing agencies the ability to determine their own risk posture. FedRAMP 20x represents our commitment to cutting through complexity, empowering innovation, and ensuring that security keeps pace with technological advancement. FedRAMP 20x will keep driving faster, smarter, and more customer-focused service for years to come.” 「FedRAMPは政府機関の重要なニーズを満たす共有サービスだ。「我々は単にプロセスを近代化するだけでなく、連邦政府のクラウドセキュリティのあり方を再構築し、各機関が自らのリスク態勢を決定する能力を提供している。FedRAMP 20xは、複雑さを克服し、イノベーションを強化し、セキュリティが技術の進歩に遅れないようにするという我々のコミットメントを表している。FedRAMP 20xは、今後何年にもわたって、より速く、よりスマートで、より顧客重視のサービスを推進し続けるだろう。
Some of the changes being made to further position the program where it belongs, at the pace of technology, include: FedRAMP20xは、今後何年にもわたり、より迅速でスマートな、より顧客重視のサービスを推進し続けるだろう:
・No federal agency sponsor needed for simple, low-impact service offerings ・シンプルで影響の少ないサービスの提供には、連邦政府機関のスポンサーは必要ない。
・No unnecessary or duplicative paperwork ・不要な書類作成や重複する書類作成がない
・Turn-key adoption for simple, cloud-native environments ・シンプルでクラウド・ネイティブな環境のためのターンキー採用
・Engineer-friendly security requirements that are easy to implement ・実装が容易なエンジニアフレンドリーなセキュリティ要件
・Authorization in weeks for most cloud offerings ・ほとんどのクラウド・オファリングの認可が数週間で完了
“As a member of the FedRAMP Board, I am incredibly excited about FedRAMP 20x,” said Chief Product Officer and Deputy Chief Information Officer, Product Delivery Service (PDS), in the Office of Information and Technology at the Department of Veterans Affairs Carrie Lee. “This transformative vision will streamline FedRAMP processes, leveraging automation and modern technologies to accelerate secure cloud adoption across federal agencies. By reducing authorization times from years to weeks and enhancing security postures through our modernization efforts, we are setting a new standard for efficiency and innovation. This initiative will lower vendor costs, increase competition, and build greater trust with industry. FedRAMP 20x is a game-changer, and I am proud to be part of this journey towards a more secure and efficient federal cloud landscape.” 「FedRAMP Boardのメンバーとして、FedRAMP 20xに大きな期待を寄せています」と、退役軍人省情報技術局のキャリー・リー最高製品責任者兼プロダクト・デリバリー・サービス(PDS)副最高情報責任者は述べた。「この変革的ビジョンは、自動化と最新技術を活用してFedRAMPプロセスを合理化し、連邦政府機関全体の安全なクラウド導入を加速する。認可にかかる時間を数年から数週間に短縮し、近代化の取り組みを通じてセキュリティ体制を強化することで、我々は効率性と革新性の新たな標準を打ち立てようとしている。このイニシアチブは、ベンダーのコストを下げ、競争を促進し、業界とのより大きな信頼を構築する。FedRAMP 20xはゲームチェンジャーであり、より安全で効率的な連邦政府のクラウド環境に向けたこの旅に参加できることを誇りに思う。
The private sector has already shared their excitement about FedRAMP 20x. FedRAMP stakeholders stated: 民間セクターはすでにFedRAMP 20xへの興奮を共有している。FedRAMP関係者は次のように述べている:
“Increased government efficiency and transformation are imperative for all agencies as they work to modernize legacy technology, streamline complex processes, and improve operations. GSA’s new approach is no exception… streamlining FedRAMP will expedite the adoption of secure, innovative technologies across government. We look forward to continuing to work closely with FedRAMP as an early adopter of FedRAMP 20x and accelerating our customers’ adoption of our solutions.” - From a Cloud Service Provider (CSP) 「レガシー・テクノロジーの近代化、複雑なプロセスの合理化、オペレーションの改善に取り組む中で、政府の効率性とガバナンスの改善はすべての政府機関にとって不可欠である。GSAの新しいアプローチも例外ではない...FedRAMPを合理化することで、政府全体で安全で革新的なテクノロジーの採用を促進することができる。我々は、FedRAMP 20xのアーリーアダプターとしてFedRAMPと緊密に協力し続け、顧客のソリューション採用を加速させることを楽しみにしている。」 - クラウド・サービス・プロバイダ(CSP)から
“We would appreciate the ability to make changes without bureaucracy.” - From a CSP "官僚的な手続きなしに変更できることを評価したい。」 - CSP から
“This is aligned to how compliance should be.” - From a CSP "コンプライアンスのあるべき姿に合致している。」 - CSPから
“The concept of government as consumers rather than approvers represents a fundamental shift.” - From an Industry Trade Association "政府が承認者ではなく、消費者であるというコンセプトは、根本的な転換を意味する。」 - 業界団体より
“It’s a step in the right direction.” - From a Third Party Assessment Association (3PAO) "正しい方向への一歩である」 - サードパーティ・アセスメント協会(3PAO)より
For more information about FedRAMP 20x, please visit fedramp.gov, read the FedRAMP 20x blog post or download the FedRAMP 20x industry engagement kit. FedRAMP 20x の詳細については、fedramp.gov を参照するか、FedRAMP 20x のブログ記事を読むか、FedRAMP 20x 業界エンゲージメント・キットをダウンロードしてください。
### ###
About GSA: GSA provides centralized procurement and shared services for the federal government. GSA manages a nationwide real estate portfolio of over 360 million rentable square feet, oversees more than $110 billion in products and services via federal contracts, and delivers technology services that serve millions of people across dozens of federal agencies. GSA’s mission is to deliver the best customer experience and value in real estate, acquisition, and technology services to the government and the American people. For more information, visit GSA.gov and follow us at @USGSA. GSA について: GSA は連邦政府に集中調達と共有サービスを提供している。GSAは、3億6,000万平方フィート以上の賃貸可能な全国的な不動産ポートフォリオを管理し、連邦政府との契約を通じて1,100億ドル以上の製品とサービスを監督し、数十の連邦政府機関にわたって数百万人にサービスを提供する技術サービスを提供している。GSAの使命は、不動産、取得、テクノロジー・サービスにおいて最高の顧客体験と価値を政府と米国民に提供することである。詳細については、GSA.govを参照し、@USGSAでフォローしてほしい。

 

 


 

 

FedRAMP

1_20250409061401

FedRAMP 20X

FEDRAMP 20X FEDRAMP 20X
FedRAMP 20x describes key goals for a new assessment process that will be designed by FedRAMP in collaboration with industry stakeholders and agency experts. Community Working Groups will drive industry innovation to provide the solution. FedRAMP 20x は、FedRAMP が業界の利害関係者や省庁の専門家と協力して設計する新たなアセスメント・プロセスの主要目標を示している。コミュニティ・ワーキング・グループが業界のイノベーションを推進し、ソリューションを提供する。
These five key goals include: これら5つの主要目標は以下の通りである:
1. Make it simple to automate the application and validation of FedRAMP security requirements. 1. FedRAMPセキュリティ要件の適用と妥当性確認を簡単に自動化する。
80%+ of requirements will have automated validation without the need to write a single word about how it works, compared to 100% of current controls requiring narrative explanations 説明的な説明を必要とする現行の管理手法が100%であるのに対して、80%以上の要件は、その仕組みについて一言も書く必要なく、自動的な妥当性確認が行われるようになる
Technical controls will make sense and match standard configuration choices 技術的な管理手法が理にかなったものとなり、標準的な構成の選択肢に合致するようになる
Industry will provide solutions and competition for varying business needs with FedRAMP aligning standards 業界は、FedRAMPが標準に整合することで、さまざまなビジネスニーズに対するソリューションと競合を提供するようになる
2. Leverage existing industry investments in security by inheriting best-in-class commercial security frameworks. 2. クラス最高の商用セキュリティ枠組みを継承することで、セキュリ ティに対する業界の既存の投資を活用する。
New documentation required for FedRAMP will be reduced to a few pages if companies provide existing security policies, change management policies, and other documentation 企業が既存のセキュリティ方針、変更管理方針、その他の文書を提供すれば、FedRAMP に新たに要求される文書は数ページに削減される
Community working groups will design optional templates that can be modified by companies to provide the foundation for remaining requirements with the approval of FedRAMP コミュニティのワーキンググループは、FedRAMP の承認を得て、企業が修正可能なオプションのテンプレートを設計し、残りの要件の基礎を提供する
Industry will provide tools to document complex technical systems by code, not narrative, that meet FedRAMP standards 業界は、FedRAMP 標準に適合する、説明ではなくコードによる複雑な技術システムを文書化するツールを提供する
3. Continuously monitor security decisions using a simple, hands-off approach. 3. シンプルなハンズオフアプローチを用いて、セキュリティの継続的なモニタリングを行う。
Industry partners will provide continuous simple standardized machine readable validation of the things that really matter 業界パートナーは、本当に重要な事柄について、シンプルで標準化された機械可読の妥当性確認を継続的に提供する
Automated enforcement and secure-by-design principles will prevent mistakes or bad decisions 自動化された実施とセキュアバイデザインの原則により、ミスや誤った決定を防止する
Community working groups will collaborate with FedRAMP to ensure a consistent approach across industry コミュニティのワーキンググループはFedRAMPと協力して、業界全体で一貫したアプローチを確保する
4. Build trust between industry and federal agencies by leaning into the direct business relationships between providers and customers. 4. プロバイダと顧客との間の直接的なビジネス関係に傾注することで、業界と連邦政府機関との間の信頼を構築する。
Cloud service providers and agencies will interact directly over established business channels to review and maintain security クラウド・サービス・プロバイダと連邦政府機関は、確立されたビジネス・チャネルを通じて直接対話し、セキュリティのレビューと維持を行う
Industry trade groups can band together to establish shared procedures that work best for them, or companies can set out alone, as long as minimum requirements set by FedRAMP are met 業界の業界団体は、FedRAMP が定める最低限の要件を満たす限り、結束して自社に最適な共有手順を確立することも、企業が単独で着手することもできる
Businesses will maintain control of their intellectual property and make their own decisions on how it can be shared 企業は知的財産の管理を維持し、その共有方法について自ら決定する
5. Enable rapid continuous innovation without artificial checkpoints that halt progress. 5. 進歩を止める人為的なチェックポイントを設けることなく、迅速で継続的なイノベーションを可能にする。
Industry will implement enforcement systems that ensure security is constantly in place; annual assessments will be replaced by simple automated checks 産業界は、セキュリティが常に確保されていることを保証する実施システムを導入する。毎年のアセスメントは、単純な自動チェックに取って代わられる
Significant changes that follow an approved established business process won’t require additional oversight 承認された確立されたビジネスプロセスに従った大幅な変更は、追加の監視を必要としない
Industry will help FedRAMP establish clear, consistent guidelines for making big changes that level the playing field between companies without ghost regulations 産業界は、FedRAMPが、幽霊のような規制なしに企業間の競争条件を平準化するような大きな変更を行うための明確で一貫したガイドラインを確立するのを支援する。

 

FedRAMP 20x Community Working Groups

 

FedRAMP 20x - Engagement

・・[PDF] FedRAMP 20x One Pager

20250409-62106

FedRAMP 20X FedRAMP 20X
FedRAMP 20x offers a cloudnative continuous security assessment that’s as simple as your cloud service offering. FedRAMP 20x は、クラウド・サービスと同じくらいシンプルな、クラウドネイティブな継続的セキュリティ・アセスメントを提供する。
FedRAMP 20x Overview FedRAMP 20x の概要
Modernization begins today. FedRAMP 2025 is a revolutionary approach to how government should engage with industry. We're removing blockers and streamlining processes for simpler, easier and cheaper cloud adoption. モダナイゼーションは今日から始まる。FedRAMP 2025は、政府が産業界とどのように関わるべきかについての画期的なアプローチである。クラウドの導入をよりシンプルに、より簡単に、より安価にするために、阻害要因を取り除き、プロセスを合理化する。
2025 Mission  2025 ミッション
FedRAMP is a government-wide program that sets the standards and policies to galvanize private innovation to create best-in-class secure cloud solutions. FedRAMPは政府全体のプログラムであり、民間のイノベーションを活性化し、クラス最高のセキュアなクラウド・ソリューションを生み出すための標準とポリシーを定める。
What’s Changing For Cloud Providers クラウドプロバイダーにとって何が変わるか
⦁ Agency sponsorship will not be necessary because authorization will be simple ⦁ 認可がシンプルになるため、行政機関のスポンサーシップは不要になる
⦁ Use existing security certifications to prove system security standards are met ⦁ システムのセキュリティ基準を満たしていることを証明するために、既存のセキュリティ認証を利用する
⦁ Choose what you want to offer and let the agency choose ⦁ 提供したいものを選択し、行政機関に選択させる
⦁ Authorizations will take weeks instead of months and years ⦁ 認可には数カ月や数年ではなく、数週間になる
⦁ Work independently with an agency after cloud offering adoption ⦁ クラウドオファーの採用後は、行政機関と独立して仕事をする
⦁ Continuous monitoring will be decentralized and will happen on your terms ⦁ 継続的な監視は非中央集権化され、顧客の条件で行われる
How Cloud Providers Will Benefit クラウドプロバイダーはどのような恩恵を受けるか
Reducing duplicative efforts and minimizing documentation saves you time and money 重複する取り組みを減らし、文書化を最小限に抑えることで、時間と費用を節約できる
Open forum for innovation to deliver cloud services for the American taxpayer 米国の納税者のためにクラウドサービスを提供するイノベーションのためのオープンフォーラム
Greater ROI for adding new services to the FedRAMP Marketplace quickly 新しいサービスをFedRAMPマーケットプレイスに迅速に追加することで、ROIがより大きくなる。

 

FedRAMP 20x Industry Engagement Kit

20250409-63156

 

FEDRAMP 20X FAQS

FedRAMP 20x FAQs FedRAMP 20x FAQs
What happens to the current FedRAMP process? 現在の FedRAMP プロセスはどうなるのか?
Cloud Service Providers and federal agencies may continue to work together to perform “sponsored” FedRAMP Agency Authorizations against traditional FedRAMP Rev5 baselines and FedRAMP will accept these authorizations until a formal end-of-life timeline is announced. This means: クラウド・サービス・プロバイダと連邦政府機関は、従来のFedRAMP Rev5ベースラインに対して「スポンサー付き」FedRAMP機関認可を実施するために協力し続けることができ、FedRAMPは、正式な終了時期が発表されるまで、これらの認可を受け入れる。これは、次のことを意味する:
The FedRAMP PMO and Board will not provide updated technical assistance or guidance for implementation of the Rev5 baselines after March 2025. FedRAMP PMO と理事会は、2025 年 3 月以降、Rev5 ベースラインの実施に関する最新の技術支援やガイダンスを提供しない。
The FedRAMP PMO will stop performing in depth “triple check” reviews of FedRAMP Rev5 packages after March 2025. Agencies will be expected to review the package in depth and make their own risk assessment without the opinion of the PMO. FedRAMP PMO は、2025 年 3 月以降、FedRAMP Rev5 パッケージの詳細な「トリプルチェック」レビューの実施を停止する。アセスメントは、PMOの意見なしに、パッケージを詳細にレビューし、独自のリスクアセスメントを行うことが期待される。
The FedRAMP PMO will halt the limited centralized continuous monitoring of former JAB-authorized FedRAMP Rev5 cloud service offerings after March 2025 and authorizing agencies will be responsible for monitoring. A Community Working Group will coordinate with industry to update this process. FedRAMP PMOは、2025年3月以降、旧JAB認可のFedRAMP Rev5クラウド・サービスの限定的な集中継続監視を停止し、作成機関が監視の責任を負う。コミュニティ・ワーキング・グループは、このプロセスを更新するために産業界と調整する。
What is FedRAMP 20x? FedRAMP 20xとは何か?
FedRAMP 20x is an initiative to partner with industry to build a cloud-native continuous security assessment that’s as simple as your cloud service offering - or as complex as needed. This new approach seeks to evaluate the outcomes of automated monitoring and enforcement of commercial security best practices to meet the minimum security requirement for federal information systems. FedRAMP 20xは、クラウド・ネイティブな継続的セキュリティ・アセスメントを構築するために、産業界と提携するイニシアティブである。この新しいアプローチは、連邦情報システムの最低セキュリティ要件を満たすために、商用セキュリティのベスト・プラクティスを自動監視・実施した結果を評価しようとするものである。
Why this and why now? なぜ今なのか?
We’ve heard your feedback that the FedRAMP authorization process is too expensive, time-consuming, and challenging. GSA is dedicated to bringing more cloud services to government while effectively managing risks. As part of the Trump-Vance transition towards increased government efficiency, we are transitioning away from costly, inefficient, manually compiled documentation and towards industry-led, data-driven security reporting. 我々は、FedRAMP認可プロセスがあまりにも高価で、時間がかかり、困難であるという意見を聞いてきた。ガバナンスは、リスクを効果的にマネジメントしながら、より多くのクラウド・サービスを政府に提供することに専念している。政府の効率化に向けたガバナンスの一環として、我々は、高価で非効率的な手作業で作成された文書から、業界主導のデータ主導のセキュリティ報告へと移行しつつある。
What are the next steps? When will FedRAMP 20x be implemented? 次のステップは何か?FedRAMP 20x はいつ実施されるのか?
Technical assistance and guidance for FedRAMP 20x will be formalized on a rolling basis as the pilot is validated by the Community Working Groups. Each piece of guidance will go through formal public comment before it is made official and open to use by industry and other agencies. FedRAMP 20x のための技術支援とガイダンスは、コミュニティ・ワーキング・グループによる試験的な妥当性確認が進むにつれて、順次正式化される予定である。各ガイダンスは、正式なパブリック・コメントを経てから、産業界や他の機関が利用できるようになる。
What about existing FedRAMP authorized cloud service offerings? 既存のFedRAMP認可クラウド・サービスはどうなるのか?
All currently authorized cloud service offerings will be designated as FedRAMP Rev. 4 or Rev. 5 Authorized until they update to a newer 2025 or higher baseline. 現在認可されているクラウド・サービスはすべて、新しい2025年またはより高いベースラインに更新されるまでは、FedRAMP Rev. 4またはRev. 5 Authorizedとして指定される。
I’m a new cloud provider. How do I get authorized today? 私は新しいプロバイダだ。どうすれば認可されるのか?
The only available route to FedRAMP authorization today is the Rev. 5 Agency Authorization path outlined on the FedRAMP website: [web] 現在FedRAMP認可を受けることができる唯一のルートは、FedRAMPのウェブサイト(web )で説明されているRev.5の認可機関パスである。
Will FedRAMP 20x remain the same in 2026, 2027, etc.? FedRAMP 20x は 2026 年、2027 年などでも変わらないのか?
FedRAMP will be continuously improved and updated on a yearly basis. FedRAMP 20x is initially focused on cloud-native software-as-a-service, deployed on an existing FedRAMP Authorized cloud service offering using entirely or primarily cloud-native services, with minimal or no third party cloud interconnections. FedRAMP は継続的に改善され、毎年更新される。FedRAMP 20xは当初、クラウド・ネイティブなSaaSに焦点を当て、既存のFedRAMP認可クラウド・サービス上に展開され、クラウド・ネイティブなサービスを完全に、あるいは主に使用し、サードパーティーのクラウド相互接続は最小限、あるいは全くないものとする。
What if I’m not able to join a Community Working Group? Can I still provide feedback? コミュニティ・ワーキンググループに参加できない場合はどうすればよいのか?それでもフィードバックを提供できるか?
Absolutely; while the Community Working Groups will work to validate initial ideas and encourage adoption, there will be an opportunity to share your feedback on any draft guidance during the formal public comment period. This approach allows room for continuous iterations before the first phase of FedRAMP 2025 launches. コミュニティ・ワーキンググループは、初期のアイデアを妥当性確認し、採用を促進するために活動するが、正式なパブリックコメント期間中に、ドラフトガイダンスに対するフィードバックを共有する機会が設けられる。このアプローチにより、FedRAMP 2025の第一段階が開始される前に、継続的な反復を行う余地が生まれる。
How will it work for a cloud service provider currently in the authorization pipeline? 現在認可パイプラインにあるクラウド・サービス・プロバイダにとってはどうなるのか?
Cloud service providers and federal agencies may continue to work together to perform “sponsored” FedRAMP Agency Authorizations against traditional FedRAMP Rev. 5 baselines and FedRAMP will accept these authorizations until a formal end-of-life timeline is announced. However, FedRAMP will not provide updated technical assistance or guidance for implementation of the Rev. 5 baselines. Agencies will be expected to review the package in depth and independently make their own risk assessment. クラウド・サービス・プロバイダと連邦政府機関は、従来のFedRAMP Rev.5ベースラインに対して「スポンサー付き」FedRAMP機関認可を実施するために協力し続けることができ、FedRAMPは正式な終了時期が発表されるまでこれらの認可を受け入れる。しかし、FedRAMP は、Rev.5 ベースラインの実施に関する最新の技術支援やガイダンスを提供しない。アセスメントは、このパッケージを詳細に検討し、独自にリスクアセスメントを行うことが期待される。
How can I be sure to get notified of FedRAMP 20x changes? FedRAMP 20x の変更の通知を確実に受け取るにはどうしたらよいか?
FedRAMP believes in transparency and open collaboration. Be sure to follow along with our progress on GitHub link and through our Change Log on fedramp.gov/changelog. FedRAMP は透明性とオープンなコラボレーションを信条としている。GitHubのリンクやfedramp.gov/changelogの変更履歴で、進捗をフォローしてほしい。
Will new cloud service providers need an agency “sponsor”? 新しいクラウド・サービス・プロバイダは「スポンサー」を必要とするのか?
FedRAMP 20x involves submitting both documentation and automated validation directly to FedRAMP before the cloud service offering is added to the FedRAMP Marketplace for hundreds of agencies to choose from. Once in the marketplace it will be up to agencies using a cloud service offering to authorize operation of the service as usual. FedRAMP 20xでは、クラウドサービスがFedRAMP Marketplace に追加され、何百もの機関から選択できるようになる前に、文書と自動妥当性確認の両方をFedRAMPに直接提出することになる。マーケットプレイスに登録された後は、クラウドサービスを利用する機関が通常通りサービスの運用を認可することになる。

 

 

・2025.03.24 Official Blog: FedRAMP in 2025

FedRAMP in 2025 FedRAMP in 2025
Last year FedRAMP underwent a major overhaul after more than a decade. The biggest change took place behind the scenes as the Program Management Office (PMO) onboarded an impressive cohort of federal technical experts for the first time. This team of federal security experts, platform and software engineers, data scientists, and communication strategists are backed by individuals with proven experience in leadership that have built cloud services and managed actual security programs. 昨年、FedRAMPは10年以上ぶりに大改革を行った。最大の変化は舞台裏で行われ、プログラム・マネジメント・オフィス(PMO)が連邦政府の技術専門家からなる素晴らしい集団を初めて迎え入れたことだ。連邦政府のセキュリティ専門家、プラットフォーム・エンジニア、ソフトウェア・エンジニア、データ・サイエンティスト、コミュニケーション・ストラテジストで構成されるこのチームは、クラウド・サービスの構築や実際のセキュリティ・プログラムの管理でリーダーシップを発揮してきた経験豊富な人材に支えられている。
It’s that foundation of expert staff that will ensure a successful transformation of FedRAMP into a streamlined, automation-driven compliance framework that accelerates secure cloud adoption across federal agencies while leveraging modern technologies to minimize bureaucracy and maximize efficiency in 2025. FedRAMPを合理化された自動化主導のコンプライアンスフレームワークへと転換させ、連邦政府機関全体で安全なクラウドの導入を加速させるとともに、最新のテクノロジーを活用して官僚主義を最小限に抑え、2025年の効率を最大化することを確実にするのは、このような専門スタッフの基盤なのだ。
Here’s what you need to know about the continued evolution of FedRAMP up front: FedRAMPの継続的な進化について、前もって知っておくべきことは以下の通りだ:
1. The existing Agency Authorization path based on FedRAMP Rev. 5 baselines is the sole active path to FedRAMP authorization. No changes to this path are planned at this time. Companies and agencies that have active investments in achieving FedRAMP authorization via this path are encouraged to evaluate the progress of FedRAMP’s efficiency improvement initiatives to make their own informed decisions. 1. FedRAMP Rev.5ベースラインに基づく既存の作成機関認可パスは、FedRAMP認可への唯一の有効なパスである。現時点では、このパスへの変更は予定されていない。このパスを通じてFedRAMP認可を取得するために積極的な投資を行っている企業や機関は、FedRAMPの効率改善イニシアチブの進捗状況を評価し、十分な情報に基づいた決定を行うことが推奨される。
2. FedRAMP will collaborate publicly with industry and other stakeholders to build and iteratively improve a new authorization process that is designed to be cloud-native and simple to automate, allowing companies to continuously and efficiently validate the underlying security of their services. This new framework, FedRAMP 20x, will be updated yearly to encourage ongoing improvements in security. 2. FedRAMPは、クラウドネイティブで自動化が簡単な新しい認可プロセスを構築し、反復的に改善するために、業界やその他の利害関係者と公に協力する。この新しい枠組みFedRAMP 20xは、セキュリティの継続的改善を促すために毎年更新される。
3. FedRAMP will not build on the old ways to consolidate resources and services that turn FedRAMP into a slow bureaucratic behemoth operating on behalf of the entire government. Instead, FedRAMP will clear the way for the development of new paths that focus on true security and eliminate the inefficiencies, making central services unnecessary. FedRAMP will set the standards and policies that enable private innovation to create the solution. 3. FedRAMPは、FedRAMPを政府全体を代表して運営する遅い官僚的巨大組織にしてしまうような、リソースとサービスを統合する古い方法を土台にするものではない。その代わりに、FedRAMPは真のセキュリティに焦点を当て、非効率性を排除し、中央サービスを不要にする新しい道を開発する道を開く。FedRAMPは、民間のイノベーションが解決策を生み出すことを可能にする標準とポリシーを設定する。
4. The new FedRAMP PMO is a much smaller team with all efforts focused on maximizing efficiency. We are now focused on clearing the agency authorization backlog and providing technical assistance and community support to set standards that enable private innovation to provide the solution. Nearly all other previously discussed work has been stopped. 4. 新しいFedRAMP PMOは、効率性を最大化することに全力を注ぐ、はるかに小規模なチームである。我々は現在、作成機関の認可の滞りを解消し、民間のイノベーションがソリューションを提供できるようにする標準を設定するための技術支援とコミュニティ支援を提供することに集中している。以前議論されたその他の作業はほぼすべて中止された。
Stay informed through definitive resources 決定的な情報源を通じて常に情報を得る
FedRAMP will continue to share information publicly through our website, working groups, speaking events, and formal requests for comment. Stakeholders are strongly encouraged to rely on these official sources for accurate, up-to-date information. Third parties discussing FedRAMP are strongly encouraged to link directly to these official resources to avoid confusion: FedRAMPは、ウェブサイト、作業部会、講演会、正式な意見要求などを通じて、引き続き情報を公開していく。関係者は、正確な最新情報については、これらの公式情報源を信頼することが強く推奨される。FedRAMPについて議論するサードパーティは、混乱を避けるため、これらの公式リソースに直接リンクすることが強く推奨される:
The FedRAMP 20x page provides information about our work to create a new authorization process, information on pilot eligibility, and next steps FedRAMP 20x のページでは、新しい認可プロセスを作成するための私たちの作業、試験的資格に関する情報、および次のステップに関する情報を提供する
The Community Working Groups page shares high-level information about our public engagement and collaboration plans コミュニティ作業部会のページでは、私たちのパブリック・エンゲージメントとコラボレーション計画に関するハイレベルな情報を共有する
The FedRAMP 20x Engagement page tracks future and past public events, press coverage, and podcast interviews FedRAMP 20x のエンゲージメンのページでは、将来および過去のパブリック・イベント、報道、ポッドキャスト・インタビューを追跡する
The FedRAMP 20x Frequently Asked Questions page tracks official answers to commonly asked questions about FedRAMP 20x FedRAMP 20x のよくある質問のページでは、FedRAMP 20x に関するよくある質問に対する公式回答を追跡する
The Changelog page tracks significant information updates all in one place 変更履歴 のページでは、重要な情報の更新を一箇所にまとめて追跡する
To view GSA’s official press announcement regarding the upcoming changes to FedRAMP, see here. FedRAMP の今後の変更に関する GSA の公式報道発表を見るには、ここを参照のこと。

 

2025.03.23 Federal News Network: GSA’s overhaul of FedRAMP contingent on automation

2025.03.20  Nextgov/FCW: FedRAMP to announce major overhaul next week

 


 

 

 NIST

OSCAL: the Open Security Controls Assessment Language

learning materials 

 


 

AMAZONのOSCAL対応の件

● AMAZON AWS - AWS Security Blog

・2022.06.30 AWS achieves the first OSCAL format system security plan submission to FedRAMP

 

AMAZON AWSが対応をしている監査プログラム...

日本語です...

コンプライアンスプログラムによる対象範囲内の AWS のサービス

 


 

 まるちゃんの情報セキュリティ気まぐれ日記

OSCAL

・2024.07.22 米国 これからはFedRampも含めて監査は自動化 (automate.fedramp.gov) !キーワードはOSCAL

・2024.05.24 米国 NIST SP 800-229 2023会計年度サイバーセキュリティとプライバシー年次報告書

・2024.04.22 内部監査人協会 意見募集 トピック別要求事項:サイバーセキュリティ (2024.04.11)

・2024.04.02 米国 FedRAMPの新しいロードマップ(2024-2025)

・2024.03.06 米国 NIST IR 8278 Rev. 1 国家オンライン情報参照(OLIR)プログラム: 概要、利点、利用方法、IR 8278A Rev. 1 国家オンライン情報参照(OLIR)プログラム: OLIR開発者のための提出ガイダンス

・2024.03.05 NIST IR 8477 文書化された標準、規制、フレームワーク、ガイドライン間の関係のマッピング: サイバーセキュリティとプライバシーの概念マッピングの開発 (2024.02.26)

・2023.08.21 米国 NIST 重要なハイテク産業における米国の競争力に関する報告書

・2021.08.06 NIST SP 800-53A Rev. 5 (Draft) 情報システムと組織におけるセキュリティとプライバシーのコントロールの評価

・2020.12.11 SP 800-53 Rev. 5 Security and Privacy Controls for Information Systems and Organizations, SP 800-53B Control Baselines for Information Systems and Organizationsの記載ミスの修正と第5版と第4版の更新分析、ISO/IEC27001とのマッピング表等の追加

・2020.07.16 JIPDEC ”米国のプライバシー保護に関する動向”

 

 

 

| | Comments (0)

2025.04.08

日本銀行 金融高度化センター設立20周年記念ワークショップ「デジタル化とわが国の金融の未来」 (2025.02.13)

こんにちは、丸山満彦です。

日本銀行金融機構局金融高度化センターが2025年7月に設立20周年を迎えることを記念し、2025年1月31日に設立20周年記念のワークショップを日本銀行本店において、オンライン・ライブ配信も併用して、開催したのですが、その資料等が公開されています...

金融業界のビジネス、DX、セキュリティ(耐量子暗号の話もあります...)等について、いろいろと参考になる内容だと思います...

 

日本銀行 - 金融システム

・2025.02.13 金融高度化センター設立20周年記念ワークショップ「デジタル化とわが国の金融の未来」を開催

 

「デジタル化とわが国の金融の未来」プログラム 
開会挨拶     日本銀行 金融機構局 参事役 金融高度化センター長 須藤直
【第1部】わが国の金融サービス・経済にとってのデジタル化の意義 
基調講演:デジタル化はわが国の金融経済にどのような変化をもたらすのか 資料1 動画 日本銀行 理事 高口博英 
【第2部】デジタル技術を活用した金融サービスの高度化・効率化について 
(1)プレゼンテーション       
① MUFGにおけるデジタル技術を活用した金融サービスの高度化・効率化 資料2   株式会社三菱UFJフィナンシャル・グループ 執行役常務 山本忠司 
② SMBCグループがデジタルで作り出す新規事業 資料3 動画 株式会社三井住友フィナンシャルグループ 執行役員 白石直樹 
③ FFGにおける企業変革のためのDX戦略 資料4   株式会社ふくおかフィナンシャルグループ 執行役員 藤井雅博 
④ 信用金庫業界におけるデータ利活用を通じた業務高度化について 資料5 動画 信金中央金庫 常務理事 神野善則
⑤ 野村グループにおけるDXの取り組み 資料6 動画 野村ホールディングス株式会社 執行役員 池田肇
⑥ 生成AIからAIエージェントの時代へ 資料7   日本マイクロソフト株式会社 業務執行役員 金子暁
(2)パネルディスカッション       
  パネリスト 山本忠司、白石直樹、藤井雅博、神野善則、池田肇、金子暁 
  モデレータ 須藤直 
【第3部】デジタル技術を活用した金融サービスの安定的な提供について 
(1)プレゼンテーション       
① 金融システムにおける暗号技術利用の変容とセキュリティ上の対応 資料8 動画 ジョージタウン大学 リサーチ・プロフェッサー松尾真一郎 
② 耐量子計算機暗号(PQC)と金融機関の対応について 資料9 動画 株式会社みずほフィナンシャルグループ グループ執行役員 寺井理 
③ 生成AI導入に向けた課題とリスク 資料10   株式会社NTTデータ 執行役員 細谷好志
④ 金融イノベーションと安定性の両立 資料11   アマゾンウェブサービスジャパン合同会社 常務執行役員 鶴田 規久
⑤ Fintech観点で考える安定化へのヒント 資料12 動画 株式会社マネーフォワード 執行役員 瀧俊雄
⑥ 金融分野におけるセキュリティの潮流 資料13   日本銀行 金融研究所 参事役 情報技術研究センター長 鈴木淳人 
(2)パネルディスカッション       
  パネリスト 松尾 真一郎、寺井理、細谷好志、鶴田規久、瀧俊雄、鈴木淳人  
  モデレータ 日本銀行 金融機構局 参事役 兼 上席考査役 徳高康弘
【第4部】まとめ       
コメント     東京大学 教授 柳川範之
閉会挨拶     日本銀行 須藤直

 

1_20250408055801

 

 

| | Comments (0)

2025.04.05

欧州委員会 B2Bデータ共有とクラウドコンピューティング契約に関する専門家グループの最終報告書 (2025.04.03)

こんにちは、丸山満彦です。

2022年から検討をされていた、B2Bデータ共有とクラウドコンピューティング契約に関する専門家グループが最終報告書を公表していますね...

データ法第41条に基づくモデル契約(Model Contractual Terms: MCT)と標準契約約款(Standard Contractual Clauses: SCC)についてのガイドで、データ法を遵守した上で、データアクセスとデータ利用を促進することを目的としていますね...

データ法第41条

Article 41; Model contractual terms and standard contractual clauses 第41条 モデル契約条項および標準契約条項
The Commission, before 12 September 2025, shall develop and recommend non-binding model contractual terms on data access and use, including terms on reasonable compensation and the protection of trade secrets, and non-binding standard contractual clauses for cloud computing contracts to assist parties in drafting and negotiating contracts with fair, reasonable and non-discriminatory contractual rights and obligations. 委員会は、2025年9月12日までに、当事者が公正、合理的かつ非差別的な契約上の権利および義務を有する契約を作成し、交渉するのを支援するため、合理的な補償および営業秘密の保護に関する条項を含む、データへのアクセスおよび使用に関する拘束力のないモデル契約条項、ならびにクラウドコンピューティング契約に関する拘束力のない標準契約約款を作成し、勧告するものとする。

 

DFFTを推進する上でも参考の一つになるでしょうね...

European Commission - Register of Commission Expert Groups - Expert Groups

Expert Group on B2B data sharing and cloud computing contracts (E03840)

最終報告書...

 ・[PDF]  ダウンロードされます...

20250405-62907

 

Introduction  序文
I. Context  I. 文脈
The Expert Group  専門家グループ
According to Article 41 of the Data Act the Commission should recommend non-binding model contractual terms on data access and use (‘MCTs’) and standard contractual clauses for cloud computing contracts (‘SCCs’). In 2022 the Commission set up the Expert group on B2B data sharing and cloud computing contracts to assist it with the development of the terms and clauses.  データ法第41条によると、欧州委員会は、データへのアクセスと使用に関する拘束力のないモデル契約条項(「MCTs」)と、クラウドコンピューティング契約の標準契約条項(「SCCs」)を推奨すべきである。2022年、欧州委員会はB2Bデータ共有および専門家グループをクラウドコンピューティング契約に関する設置し、契約条項の策定を支援している。
The expert group consists of 17 members acting in individual capacity, mostly lawyers, but also other practitioners and academics with experience in contracts for sharing data between companies and in cloud computing services. The experts started by identifying together the issues to be covered by the models, collecting use-cases, industry standards and similar or related contractual models and exchanging based on their own experience.  専門家グループは、弁護士を中心に、企業間のデータ共有契約やクラウドコンピューティング・サービスの経験を持つ実務家や学識経験者など、個人の立場で活動する17名のメンバーで構成されている。専門家たちはまず、モデルでカバーすべき問題を特定し、ユースケース、業界標準、類似または関連する契約モデルを収集し、自らの経験に基づいて意見交換を行った。
The first meeting of the Expert Group was held in September 2022. Until the end of March 2025 the Expert Group met 19 times in an official format, chaired by the Commission services that ensured the secretariat of the group. In between these meetings, the experts worked in smaller drafting group for each clause, which was afterwards discussed during the plenary meetings.  専門家グループの初会合は2022年9月に開催された。2025年3月末まで、専門家グループは公式形式で19回会合を開き、その議長は、グループの事務局を確保した欧州委員会のサービスが務めた。これらの会合の合間に、専門家たちは条項ごとに少人数のドラフトグループに分かれて作業を行い、その後、本会議で議論が行われた。
The sub-group  サブグループ
The Commission also set up a sub-group comprising of companies, European and nation-wide organisations active in data sharing and cloud computing. The sub-group gave feedback to the experts throughout the drafting process. All draft terms and clauses that the Expert Group prepared were subject to consultation with the sub-group, sometimes several times, and their input was then considered by the experts in view of improving the texts.  また、欧州委員会は、データ共有とクラウド・コンピューティングに積極的な企業、欧州組織、国家組織からなるサブグループを設置した。このサブグループは、ドラフト作成過程を通じて専門家グループにフィードバックを行った。専門家グループが作成したすべての条項のドラフトは、サブグループとの協議の対象となり、時には数回にわたって協議が行われた。
Consultation activities  コンサルテーション活動
In addition to the feedback received via the sub-group, the draft terms and clauses were also subject to consultation via a testing exercise organised during summer 2024. 12 companies provided feedback on the MCTs and 14 on the SCCs. The experts debated the results and made changes to the drafts where needed.  小委員会を通じて寄せられたフィードバックに加え、2024年夏に実施されたテストを通じて、ドラフト条項が協議の対象となった。MCTについては12社、SCCについては14社がフィードバックを提供した。専門家はその結果を討議し、必要に応じてドラフトに変更を加えた。
In November and December 2024 the Commission organised a series of six public webinars, where several members of the Expert Group presented the MCTs and SCCs. Between 200 and 300 participants took part in each webinar and participants had the opportunity to express opinions, drafting suggestions and elaborate with examples. Following these webinars, the experts further improved the drafts.  2024年11月と12月、欧州委員会は6回にわたる公開ウェビナーを開催し、専門家グループの複数のメンバーがMCTsとSCCsについて説明した。各ウェビナーには200人から300人の参加者があり、参加者は意見を述べたり、ドラフトに関する提案をしたり、例を挙げて詳しく説明したりする機会を得た。これらのウェビナーの後、専門家たちはさらにドラフトを改善した。
II. Nature and use of the clauses  II. 条項の性質と使い方
The MCTs and SCCs are non-binding, voluntary and have been drafted so they can be adapted by the parties according to their contractual needs. However, the parties need to consider that these models were drafted to be in line with the rights and obligations provided by the Data Act and were also designed to be coherent with each other. The various models include information about their voluntary nature and warnings to the parties to consider the legal consequences when making changes. They are accompanied by an introduction with explanations and instructions about how best to use them.  MCTsとSCCsは拘束力を持たず、任意であり、当事者が契約上の必要性に応じて適応できるようドラフ トされている。しかし、当事者は、これらのモデルがデータ法によって規定された権利と義務に沿ったものであり、また相互に首尾一貫したものとなるように作成されていることを考慮する必要がある。様々なモデルには、その自発的な性質に関する情報と、変更を加える際には法的な影響を考慮するよう当事者に警告する内容が含まれている。また、序文には、その最適な使用方法についての説明と指示が付されている。
The model terms and clauses were drafted mainly for business-to-business relations. However, they can be used also in relations between business and consumers but, in that case, additional provisions would need to be added to bring the contract into compliance with mandatory consumer protection rules (e.g. the right of withdrawal of the consumer in case of contracts concluded online/at a distance).  モデル約款は、主に企業間関係を想定してドラフトされた。しかし、企業と消費者の関係でも使用することができるが、その場合、契約を強制的な消費者保護規則(例えば、オンライン/遠距離で締結された契約の場合の消費者の撤回権)に適合させるための追加条項を加える必要がある。
III. MCTs  III. モデル契約条項(MCT)
The experts analysed chapters II-IV of the Data Act and drafted the following MCTs to help implement these provisions:  専門家はデータ法第2章から第4章を分析し、これらの規定を実施するために以下のMCTを起草した:
• Data Holder to User  • データ保有者からユーザーへ
• User to Data Recipient  • ユーザーからデータ取得者へ
• Data Holder to Data Recipient  • データ保有者からデータ取得者へ
• Data Sharer to Data Recipient  • データ共有者からデータ取得者へ
IV. SCCs  IV. 標準契約約款(SCC)
Article 41 of the Data Act refers to the need for non-binding standard contractual clauses for cloud computing contracts to assist parties in drafting and negotiating contracts with fair, reasonable and nondiscriminatory contractual rights and obligations.  データ法第41条は、クラウド・コンピューティング契約における拘束力のない標準契約条項の必要性について言及しており、当事者が公正、合理的かつ差別のない契約上の権利と義務を伴う契約を作成し、交渉することを支援するものである。
Following this approach, the experts have considered the provisions of chapters VI and VIII and identified several contractual issues that are relevant for both the practical effectiveness of the newly introduced customer’s right to switch and exit as well as for the fairness and contractual balance between parties. As opposed to the approach for the MCTs, where the Expert Group drafted full contracts, the Expert Group drafted six standard clauses to cover the main contractual issues identified for the cloud computing contracts and one general clause. These clauses are meant to be inserted by the parties into the data processing services agreements.  このアプローチに従い、専門家グループは、第6章と第8章の規定を検討し、新たに導入される顧客の切替・退出権の実務上の有効性と、当事者間の公平性と契約上のバランスの両方に関連するいくつかの契約上の問題を特定した。専門家グループが完全な契約をドラフトしたMCTのアプローチとは対照的に、専門家グループは、クラウド・コンピューティング契約について特定された主な契約上の問題をカバーする6つの標準条項と1つの一般条項をドラフトした。これらの条項は、データ処理サービス契約に当事者によって挿入されることを意図している。
1. General  1. 一般
2. Switching & Exit  2. 移転と終了
3. Termination  3. 終了
4. Security & Business continuity  4. セキュリティと事業継続
5. Non-dispersion  5. 非分散
6. Liability  6. 責任
7. Non-Amendment  7. 非修正
V. How to use the MCTs and SCCs  V. MCTと使い方
When considering using these models, the parties should first read the explanations and examples included to facilitate understanding and to reflect on the most common business situations. Some models contain several options for a given term/clause.  これらのモデルの使用を検討する場合、当事者はまず、理解を容易にし、最も一般的なビジネ ス状況を反映させるために、含まれている説明と例を読むべきである。モデルによっては、ある用語・条項について複数の選択肢が含まれているものもある。
The parties should reflect on their specific needs, business relation and interests to identify the right option that best suits their contract. Usually, such options are accompanied by guidance and examples to explain the differences between them and when they are most likely to be appropriate.  当事者は、自分たちの契約に最も適した適切な選択肢を特定するために、自分たちの具体的なニーズ、ビジネス関係、利害を熟考すべきである。通常、このようなオプションには、それぞれの違いや、どのような場合に最も適切となるかを説明するガイダンスや例が添えられている。
MODEL CONTRACTUAL TERMS  モデル契約
I. Purpose of the model contractual terms  I. モデル契約目的
The model contractual terms (MCTs) set out in the Annexes have been developed to help parties draft and negotiate contracts for access to and use of data (personal and non-personal). These models aim to ensure fair, reasonable and non-discriminatory contractual rights and obligations, including reasonable compensation and the protection of trade secrets.  附属書に記載されているモデル契約条項(MCTs)は、データ(個人および非個人)へのアクセスお よび使用に関する契約のドラフトおよび交渉に役立つよう作成されている。これらのモデルは、合理的な補償や企業秘密の保護など、公正、合理的かつ非差別的な契約上の権利と義務を確保することを目的としている。
These MCTs were developed by a group of experts set up by the European Commission.  これらのMCTは、欧州委員会が設置した専門家グループによって開発された。
These models are non-binding and voluntary in nature. The parties to a contract can complement and adapt the MCTs set out in the Annexes to their individual needs and to specific EU and Member State law, where such specific law applies.  これらのモデルは拘束力を持たず、自主的なものである。契約の当事者は、附属書に記載されたMCTを補完し、個々のニーズや、特定のEU法および加盟国法が適用される場合には、それに適合させることができる。
Passages marked in grey and italics indicate that details need to be filled in by the parties.  グレーとイタリックで示された箇所は、当事者が詳細を記入する必要があることを示している。
Passages marked with [OPTION] or (if applicable) may be appropriate or not, depending on the individual situation and preferences; note that also many terms not marked in this way may be derogated from by the parties.  OPTION]または(該当する場合)でマークされた箇所は、個々の状況や嗜好によって適切であったりそうでなかったりする。(注)このようなマークが付されていない条項の多くも、当事者によって適用除外とされる可能性がある ことに留意されたい。
II. Which MCTs to use for which type of data sharing  II. どのタイプのデータどのMCTを使うか
The model contractual terms in Annex I have been designed for contracts between a data holder and a user of a connected product or related service, where the data holder wishes to use data generated using the product/service.  附属書 I のモデル契約条項は、データ保有者が製品/サービスを使用して生成されたデータ の使用を希望する場合、データ保有者と接続製品または関連サービスのユーザーとの間の 契約のために設計されている。
The model contractual terms in Annex II have been designed for contracts between a user of a connected product or related service and a third party data recipient, where the user requests a data holder to make data available to a data recipient under Article 5 of the Data Act.  附属書Ⅱのモデル契約条項は、コネクテッド製品または関連サービスのユーザーとサードパーティ のデータ取得者との間の契約のために作成された。
The model contractual terms in Annex III have been designed for contracts between a data holder and a third party data recipient who is a business, where a data holder is obliged (under Article 5 of the Data Act) to make data available to a recipient when requested to do so by a user of the product. They may also be used with appropriate modifications where a data holder is obliged to make data available to a third party data recipient under other EU law or national legislation adopted in accordance with EU law.  附属書 III のモデル契約条項は、データ保有者が(データ法第 5 条に基づき)製品のユー ザーから要求された場合にデータを取得者に提供する義務を負う場合の、データ保有者と 企業であるサードパーティ・データ受領者との間の契約用に作成された。また、データ保有者が他の EU 法または EU 法に従って採択された国内法に基づいてサードパーティーのデータ受領者にデータを提供する義務を負う場合にも、適切な修正を加えて使用することができる。
The model contractual terms in Annex IV have been designed for contracts between a data sharer and a data recipient where the data sharer wishes to make data available to a data recipient voluntarily and independent of any request by a user or similar party.  附属書IVのモデル契約条項は、データ共有者がユーザーまたは同様の当事者による 要求とは無関係に、自発的にデータ取得者がデータを利用できるようにすることを希望 する場合の、データ共有者とデータ取得者との間の契約のために作成されている。
The parties should identify their own situation by reference to the different scenarios explained above and then use the relevant MCTs.  当事者は、上記で説明したさまざまなシナリオを参考に自らの状況を特定し、関連するMCTを使用すべきである。
III. Legal value of the MCTs in relation to Data Act and other applicable law  III. データ法およびその他の適用法に関するMCTの法的価値
The MCTs seek to ensure compliance with EU and Member State law, in particular with the Data Act[1], the Data Governance Act[2] and the Trade Secrets Directive[3]. Use of the model contractual terms by contracting parties does not affect any of the rights and obligations they have under the Data Act or under other EU law or Member State law adopted in accordance with EU law, including obligations of the controller and the rights of data subjects under the General Data Protection Regulation (GDPR).[4]  MCTsは、EU法および加盟国法、特にデータ法([1] )、データガバナンス法([2] )、営業秘密指令([3] )への準拠を保証するものである。契約当事者によるモデル契約条項の使用は、データ法に基づき、またはEU法に基づき採択された他のEU法もしくは加盟国の法律に基づき当事者が有する権利および義務(管理者の義務およびデータ保護規則(GDPR)に基づくデータ対象者の権利を含む)に影響を及ぼすものではない。[4]
The parties should pay particular attention when sharing of data concerns personal data or mixed datasets. Parties are advised to refer to the "Guidance on the Regulation on a framework for the free flow of non-personal data in the European Union” (COM/2019/250 final), which addresses these notions. Especially in relation to mixed datasets, it clarifies that “if the non-personal data part and the personal data parts are ‘inextricably linked’, the data protection rights and obligations stemming from the GDPR fully apply to the whole mixed dataset, also when personal data represent only a small part of the dataset”. In particular, data holders shall make available personal data to users or third parties, as mandated under the Data Act, but only as long as there is a legal basis in accordance with the GDPR.  データの共有が個人データや混合データセットに関係する場合、当事者は特に注意を払うべきである。当事者は、これらの概念に言及している「欧州連合における非個人データの自由な流通の枠組みに関する規則に関するガイダンス」(COM/2019/250 final)を参照することが勧められる。特に混合データセットに関しては、「非個人データ部分と個人データ部分が『密接不可分に結びついている』場合、個人データがデータセットのごく一部に過ぎない場合でも、GDPRに由来するデータ保護の権利と義務は混合データセット全体に完全に適用される」と明確にしている。特に、データ保有者は、データ法の定めに従って、個人データを利用者またはサードパーティに提供しなければならないが、それはGDPRに従った法的根拠がある場合に限られる。
As for any contract, use of the model contractual terms by parties does not prevent a competent court or tribunal or any competent administrative authority from setting aside the contract or particular terms thereof for non-compliance with EU or Member State law.  あらゆる契約について、当事者がモデル契約条項を使用することは、管轄の裁判所や法廷、または管轄の行政当局が、EU法または加盟国の法律に準拠していないとして、契約またはその特定の条項を無効とすることを妨げるものではない。
IV. Who are the Parties to the model contractual terms  IV. モデル契約当事者は誰か
The contractual parties are defined at the beginning of each MCTs, following, where relevant the terms defined by the Data Act: data holder, user and data recipient.  契約当事者は、データ法によって定義された用語(データ保有者、データ利用者、データ取得者)に従い、それぞれのMCTの冒頭で定義される。
For the voluntary data sharing MCT, the experts chose to use the term ‘data sharer’ to take into account the multiple scenarios that are possible here.  自主的なデータ共有MCTについては、専門家はここで考えられる複数のシナリオを考慮し、「データ共有者」という用語を使うことにした。
For more information on the various roles and their definition, please consult the Commission’s FAQ document on the Data Act.  様々な役割とその定義の詳細については、データ法に関する文書を欧州委員会のFAQ参照されたい。
V. List of annexes and appendixes  V. 附属書および附属書リスト
Annex I: Model Contractual Terms for contracts on data access and use between data holders and users of connected products and related services  附属書 I. データ保有者とコネクティッド製品および関連利用者との間のデータアクセスおよび利用に関する契約のためのモデル契約条項
• Appendix 1 contains details of the data covered by this contract and of access arrangements  • 附属書1には、本契約の対象となるデータの詳細およびアクセスの取り決めが記載されている。
• Appendix 2 contains the form for an access request by the User  • 附属書2には、利用者によるアクセス要求の書式が記載されている。
• Appendix 3 contains the form for a request to make data available to a third party  • 附属書3には、サードパーティへのデータ提供依頼の書式が含まれている。
• Appendix 4 contains details of measures for the protection of trade secrets  • 附属書4には、営業秘密の保護措置の詳細が記載されている。
• Appendix 5 contains details on sharing data with third parties  • 附属書5には、サードパーティとのデータ共有に関する詳細が記載されている。
• Appendix 6 contains details of protection measures  • 附属書6には、保護措置の詳細が記載されている。
• Appendix 7 contains details on compensation of the User  • 附属書7には、ユーザーの補償に関する詳細が記載されている。
• Appendix 8 contains details on penalties  • 附属書8には罰則の詳細が記載されている。
• Appendix 9 contains documentation on ownership of the Product or contractual rights to use the Product or Related services  • 附属書9には、本製品の所有権または本製品もしくは関連サービスを使用する契約上の権利に関する文書が含まれる。
Annex II: Model Contractual Terms for contracts between Users and Data Recipients  附属書Ⅱ:利用者とデータ取得者間の契約に関するモデル契約条項
• Appendix 1 contains a description of the Data  • 附属書1には、データの説明が記載されている。
• Appendix 2 lists the protective measures to be taken by the Data Recipient  • 附属書2には、データ取得者が講ずべき保護措置が記載されている。
• Appendix 3 contains information on sharing the Data with third parties by the Data Recipient  • 附属書3には、データ取得者による第三者とのデータ共有に関する情報が含まれる。
• Appendix 4 contains information on compensation to the User for the Data Recipient’s use and sharing of the Data  • 附属書4には、データ取得者によるデータの使用と共有に対する利用者への補償 に関する情報が含まれている。
• Appendix 5 contains documentation on ownership of the Product or contractual rights to use the Product or Related services  • 附属書5には、本製品の所有権または本製品もしくは関連サービスを使用する契約上の権利に関する文書が含まれる。
Annex III: Model Contractual Terms for contracts between data holders and data recipients on making data available at the request of users of connected products and related services  附属書III:コネクテッド製品および関連サービスの利用者の要求に応じてデータを利用可能にすることに関する、データ保有者とデータ取得者間の契約のためのモデル契約条項
• Appendix 1 contains evidence on the request and, if applicable, any mandate  • 附属書1には、要請に関する証拠と、該当する場合は委任事項が記載されている。
• Appendix 2 contains details of the Data covered by the contract and access arrangements  • 附属書2には、契約の対象となるデータの詳細とアクセスの取り決めが記載されている。
• Appendix 3 contains details of the calculation of compensation, including the potential status of the Data Recipient as an SME  • 附属書3には、データ取得者が中小企業である可能性を含む補償金の計算の詳細が記載されている。
• Appendix 4 contains details of measures for the protection of trade secrets  • 附属書4には、営業秘密の保護措置の詳細が記載されている。
• Appendix 5 contains details on penalties  • 附属書5には罰則の詳細が記載されている。
Annex IV: Model Contractual Terms for contracts for voluntary sharing of data between business Data Sharers and Data Recipients  附属書IV:データ共有者とデータ取得者間の自発的なデータ共有に関する契約のためのモデル契約条項
• Appendix 1 contains a description of the Data  • 附属書1には、データの説明が記載されている。
• Appendix 2 contains further details regarding Data covered by a regime requiring specific measures  • 附属書2には、特定の措置を必要とする制度の対象となるデータに関する詳細が記載されている。
• Appendix 3 contains details on Personal Data and respective obligations of the Parties  • 附属書3には、個人データおよび両当事者の義務に関する詳細が記載されている。
• Appendix 4 contains applicable security measures for the sharing of Data  • 附属書4には、データ共有に適用されるセキュリティ対策が記載されている。
[1] Regulation (EU) 2023/2854 of the European Parliament and of the Council of 13 December 2023 on harmonised rules on fair access and use and amending Regulation (EU) 2027/2394 and Directive (EU) 2020/1828 (Data Act) (OJ L, 2023/2854, 22.12.2023). [1]2023年12月13日付欧州議会および理事会規則(EU)2023/2854は、公正なアクセスおよび利用に関する調和された規則であり、規則(EU)2027/2394および指令(EU)2020/1828(データ法)を改正するものである(OJ L, 2023/2854, 22.12.2023)。
[2] Regulation (EU) 2022/868 of the European Parliament and of the Council of 30 May 2022 on European data governance and amending Regulation (EU) 2018/1724 (Data Governance Act) (OJ L 152, 3.6.2022, p. 1) [2]欧州データガバナンスおよび規則(EU)2018/1724(データガバナンス法)の改正に関する2022年5月30日付欧州議会および理事会規則(EU)2022/868(OJ L 152, 3.6.2022, p. 1)
[3] Directive (EU) 2016/943 of the European Parliament and of the Council of 8 June 2016 on the protection of undisclosed know-how and business information (trade secrets) against their unlawful acquisition, use and disclosure (OJ L 157, 15.6.2016, p. 1) [3]2016年6月8日付欧州議会及び理事会指令(EU) 2016/943 未公表のノウハウ及び営業情報(営業秘密)の違法な取得、使用及び開示に対する保護に関する指令(OJ L 157, 15.6.2016, p. 1)
[4] Regulation EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation) (OJ L 119/1, 4.5.2016, p. 1–88). [4]個人データの処理に関する自然人の保護及び当該データの自由な移動に関する2016年4月27日付欧州議会及び理事会規則(EU)2016/679、並びに指令95/46/EC(一般データ保護規則)の廃止(OJ L 119/1, 4.5.2016, p. 1-88)。

 

STANDARD CONTRACTUAL CLAUSES (SCCs)  標準契約条項(SCC)
(a) Purpose of the standard contractual clauses   (a) 標準契約条項の目的
Article 41 of the Data Act requires the Commission to develop and recommend non-binding standard contractual clauses for cloud computing and other data processing services contracts (‘SCCs’) to assist parties to draft and negotiate contracts with fair, reasonable and non-discriminatory rights and obligations.  ドラフト法第41条は、委員会に対し、クラウドコンピューティングおよびその他のデータ処理サービス契約に関する拘束力のない標準契約条項(「SCC」)を策定し、推奨するよう求めている。
The SCCs, proposed by the Expert Group are intended as best practice guidance to assist the contractual implementation of the rights and obligations stemming from the Data Act and their use is fully voluntary. They are a set of modular clauses which largely complement each other but can also be used separately. The parties to a contract can complement and adapt these SCCs to their individual contractual needs, including with regard to specific EU and Member States’ law, where such specific law applies. The clauses are meant to be inserted by the parties into their data processing services agreements.   専門家グループによって提案されたSCCsは、データ法に由来する権利と義務の契約上の履行を補助するベストプラクティスガイダンスとして意図されており、その使用は完全に任意である。SCCsはモジュール式の条項の集合であり、相互に補完し合うものであるが、個別に使用することも可能である。契約当事者は、EUおよび加盟国の特定の法律が適用される場合、その法律も含め、個々の契約上の必要性に応じてこれらのSCCを補完し、適応させることができる。この条項は、当事者がデータ処理サービス契約に挿入するものである。
(b) For whom?  (b) 誰のためか?
The SCCs aim to help any Customer or Provider. This includes any Customer of any data processing service, whether public, private or otherwise, small, midsized or large.   SCCは、あらゆる顧客またはプロバイダを支援することを目的とする。これには、公的、私的、その他、小規模、中規模、大 規模を問わず、あらゆるデータ処理サービスの顧客が含まれる。
The SCCs aim to assist in particular companies that may not have sufficient experience or resources needed to draft and negotiate contracts with fair, reasonable and non-discriminatory contractual terms.   SCCは、公正で合理的かつ非差別的な契約条件のドラフトや交渉に必要な十分な経験やリソースを持たない企業を特に支援することを目的としている。
A. SCCs General with Annex Definitions [link];   A. SCCs 全般と附属書 定義 [link]; 
B. SCCs Switching & Exit [link], with Annex Switching & Exit Plan [link];  B. SCCs 移行と終了[link], 附属書 移行と終了計画 [link]
C. SCCs Termination [link];   C. SCCs の終了 [link]; 
D. SCCs Security & Business Continuity [link];   D. SCCs セキュリティと事業継続 [link];
E. SCCs Non-Dispersion [link];    E. SCCs 非分散 [link]; 
F. SCCs Liability [link], and;  F. SCCs 責任 [link], and;
G. SCCs Non-Amendment [link].   G. SCCの不同意 [link].
Below, there is a brief overview of the specific SCCs. Detailed explanations are included in the respective SCCs both in their introduction and as Info Points, listed at the end of each set of SCCs.   以下に、各SCCの簡単な概要を示す。詳細な説明は、それぞれのSCCの序文と、SCCの各セットの最後に記載されているインフォポイントの両方に含まれている。
While it is recommended to use the whole set of SCCs, it is important to note that these SCCs do not constitute the entire agreement for data processing services that would apply between a Customer and a Provider. Further topics not addressed by the Data Act and the SCCs need to be part of such an agreement.   SCCの全セットを使用することが推奨されるが、これらのSCCは顧客とプロバイ ダの間で適用されるデータ処理サービスに関する契約全体を構成するものではない ことに注意することが重要である。データ法およびSCCsで扱われていないさらな るトピックも、そのような契約の一部とする必要がある。
In line with the definition of a data processing service in the Data Act, these SCCs are applicable to all cloud service models, including IaaS, PaaS, and SaaS.  データ法におけるデータ処理サービスの定義に沿って、これらのSCCはIaaS、PaaS、SaaSを含むすべてのクラウドサービスモデルに適用される。
(c) What do the SCCs consist of?   (c) SCCは何で構成されているか?
 SCCs General   SCC全般
The SCCs General provide for a generic structure and elements, which are common to other, topicspecific SCCs. It also includes other relevant topics that generally need to be part of an agreement for data processing services. Furthermore, these SCCs comprise of the main definitions used across the other SCCs.   SCCs Generalは、他のトピック別SCCに共通する、一般的な構造と要素を提供する。また、一般的にデータ処理サービスに関する契約の一部となる必要のある、その他の関連トピックも含まれる。さらに、これらのSCCsは、他のSCCsで使用される主な定義で構成される。
 SCCs Switching & Exit, with Annex Exit Plan   SCCs 移行と終了と附属書 移行と終了計画
The Data Act ensures that any obstacles to switching are removed: technical, pre-commercial, commercial, contractual, organisational ones.  データ法は、技術的、商業前、商業的、契約的、組織的なものなど、乗り換えの障害となるものを確実に取り除くものである。
These SCCs translates the new rights and obligations introduced by the Data Act as regards the switching and exit process into contractual terms. The SCCs set out the switching process when moving from one Provider to another as well as from the cloud to on-premise infrastructure. They cover the process and timelines for switching, responsibilities of the parties involved, assistance to be made available to the Customer (and any third parties authorised by the Customer), and further related topics mandatory under the Data Act.   これらのSCCは、データ法によって導入された、切り替えと終了プロセスに関する新たな権利と義務を契約条件に翻訳したものである。SCCは、プロバイダから別のプロバイダへ、またクラウドからオンプレミスのインフラへ移行する際の切り替えプロセスを定めている。SCCには、切り替えのプロセスとスケジュール、関係者の責任、顧客(および顧客が承認したサードパーティ)が利用できる支援、データ法で義務付けられているその他の関連事項が記載されている。
SCCs Termination   SCCの終了
These SCCs cover the service contract termination process. They are directly linked to the SCCs Switching & Exit and provide for the various termination possibilities and related conditions. The data processing agreement is usually terminated (in full or partially) once the switching process has been concluded successfully or if the Customer does not wish to switch to another Provider but requires all data to be erased and such erasure is successful. Some additional options are included in these SCCs to assist the parties under scenarios not covered in the Data Act.   これらのSCCは、サービス契約の終了プロセスをカバーする。これらのSCCは、SCC「切替および終了」に直接リンクされ、様々な終了の可能 性および関連条件を規定する。通常、データ処理契約は、切り替えプロセスが成功裏に完了した場合、または顧客 が他のプロバイダへの切り替えを希望せず、全データの消去を要求し、その消去 が成功した場合に、(全部または部分的に)終了する。これらのSCCには、データ法に規定されていないシナリオの下で、両当事者を支援す るための追加オプションがいくつか含まれている。
SCCs Security & Business Continuity   SCC セキュリティと事業継続
These SCCs are based on the requirement of the Data Act that throughout the switching process, a high level of security should be maintained and that the Provider should act with due care to maintain business continuity and ensure the provision of the functions or services under the Agreement, as well as to provide clear information concerning known risks to continuity.  これらのSCCは、スイッチングプロセス全体を通じて、高水準のセキュリティが維持されるべきであり、プロバイダは事業継続性を維持し、本契約に基づく機能またはサービスの提供を確保するために十分な注意を払って行動すべきであり、また、継続性に対する既知のリスクに関する明確な情報を提供すべきであるというデータ法の要件に基づいている。
While the topics of security and business continuity is generally already included in the data processing agreement and other contractual documentations, however, these SCCs focus on the importance of security and business continuity with relation to the switching and exit process as per the rights and obligations introduced by the Data Act.   一般的に、セキュリティと事業継続のトピックは、データ処理契約書やその他の契約文書にすでに含まれているが、これらのSCCは、データ法によって導入された権利と義務に従って、切り替えと終了プロセスに関するセキュリティと事業継続の重要性に焦点を当てている。
 SCCs Non-Dispersion    SCC 非分散
The SCCs Non-Dispersion provide for information and communication symmetry. Data processing agreements often comprise of various ancillary documents, which refer to additional material and information, which can be widely dispersed across different repositories. This makes it difficult for prospective Customers to find all the information they need to assess the parties’ contractual and legal obligations and their implications. These SCCs are, therefore, intended to help easy and practical access to relevant updated information, documents, materials and contact details by both Customer and Provider.   SCCsのNon-Dispersionは、情報とコミュニケーションの対称性を提供する。データ処理契約は多くの場合、追加資料や情報を参照する様々な付属文書で構成され、 それらは異なる保管場所に広く分散している可能性がある。このため、顧客となる見込みのある者は、当事者の契約上・法律上の義務やその意味を評価するために必要な情報をすべて見つけることが困難となる。従って、このSCCは、顧客とプロバイダの双方が、関連する最新の情報、 文書、資料、連絡先詳細に容易かつ実用的にアクセスできるようにすることを目的と している。
SCCs Liability   SCC 責任
These SCCs Liability assist the parties in defining balanced and mutually appropriate liability terms. In particular, small and mid-size companies and other organisations may find these SCCs helpful when they negotiate and/or draft liability provisions in their Agreements. The clauses also point at some issues that organisations, irrespective of their size, may not have thought of.  これらのSCCは、当事者がバランスの取れた、相互に適切な賠償責任条項を定義する際に役立つ。特に、中小企業やその他の組織は、契約書の賠償責任条項の交渉やドラフトを行う際に、このSCCsが役に立つだろう。また、規模の大小にかかわらず、組織が思いつかないような問題も指摘されている。
While the Expert Group acknowledges that a clause on liability is usually included in the general Agreement between the Customer and the Provider, the experts believe it is useful to include optional provisions on liability in the SCCs as they reflect the spirit of the Data Act of the importance of fair, reasonable and non-discriminatory rights and obligations, including those related to the cloud computing contracts and the switching process.  専門家グループは、責任に関する条項は通常、顧客とプロバイダ間の一般契約に含まれることを認めつつも、クラウドコンピューティング契約や切り替えプロセスに関連するものを含め、公正、合理的かつ非差別的な権利と義務の重要性というデータ法の精神を反映するものであるため、SCCに責任に関するオプション条項を含めることは有用であると考える。
 SCCs Non-Amendment   SCC 不参加
It is important that the parties can rely on the rights and obligations they agreed to contractually and that these rights and obligations are not changed or otherwise amended unilaterally, unless under clearly and mutually agreed conditions.   重要なのは、当事者が契約上合意した権利と義務に依拠することができ、明確かつ相互に合意された条件下でない限り、これらの権利と義務が一方的に変更されたり、その他の形で修正されたりしないことである。
The SCCs Non-Amendment are intended to give more clarity and confidence to the parties when agreeing on terms pertaining to unilateral changes. The SCCs address possible contractual arrangements to help ensure that unilateral changes are not detrimental to the interests of either party.   SCCは、一方的な変更に関連する条件について合意する際、当事者により明確さと信頼を与えることを意図している。SCCsは、一方的な変更がどちらかの当事者の利益を損なわないようにするための、可能な契約上の取り決めについて述べている。
(d) How to use the SCCs?   (d) SCCをどのように使うか?
It is recommended to start with and from the SCCs General, in which the complete set of SCCs is mentioned. The complete set of SCCs was developed to be consistent with and mutually reinforcing each other. Although not necessary, it is recommended to use the entire set of SCCs.  SCCの全セットが記載されている『SCCs General』から始めることをお勧めする。SCCsの全セットは、互いに一貫性があり、相互に補強し合うように開発された。必要ではないが、SCCの全セットを使用することを推奨する。
Each of the SCCs first explains the chosen approach or specific situations covered. It is recommended to familiarise oneself with these explanations.   各SCCはまず、選択されたアプローチや特定の状況について説明している。これらの説明をよく理解することをお勧めする。
For ease of reference, the SCCs generally contain cross-references that link certain provisions of the SCCs to relevant provisions in the Data Act or to other SCCs.   参照を容易にするため、一般規定には、SCCの特定の規定をデータ法または他のSCCの関連規定にリンクさせる相互参照が含まれている。
The parties can discuss, negotiate, and agree to cover a multitude of services in one, single agreement, which may be called master services agreement or otherwise (‘Agreement’). This is another reason why it is recommended to add the SCCs to the Agreement. If it covers all services, and if the Customer may decide to switch one or a given number of services to a new, designated Provider, the Agreement will continue to be in force for the other services.   両当事者は、マスターサービス契約などと呼ばれる1つの契約書(「契約書」)で、多数のサービスをカバーすることを協議、交渉、合意することができる。SCCを契約書に追加することが推奨されるもう1つの理由はこれである。SCCsがすべてのサービスをカバーし、顧客が1つまたは指定された数のサービスを新しい指定されたプロバイダに切り替えることを決定する可能性がある場合、本契約は他のサービスについても効力を継続する。
 Please note that, if other definitions than those proposed in the SCCs are used, the SCCs have to be adapted accordingly. This is just one example of adaptations that are possible while still using the complete set, or a part of the SCCs.  SCCsで提案されている定義以外の定義を使用する場合、SCCsをそれに合わせて適応させなければならないことに注意されたい。これは、SCCs一式またはSCCsの一部を使用しながらも可能な適応の一例に過ぎない。
If you intend to use these SCCs only partially or to make changes to them, you should carefully consider how this might affect the contractual rights and obligations.   これらのSCCを部分的にのみ使用する場合、またはSCCに変更を加える場合、契約上の権利および義務にどのような影響が及ぶかを慎重に検討する必要がある。
Legal and other professional advice is always recommended in such situations!   このような状況では、常に法律その他の専門家の助言を受けることを推奨する!
In any case, the s the users of these SCCs, whether Customers, Providers or others, decide  how to use the SCCs, entirely or partially, or not to use them  .   いずれにせよ、SCCをどのように使用するかは、顧客、プロバイダ、その他を問わず、SCCの使用者が、SCCの全部または一部を使用するか、あるいはSCCを使用しないかを決定する。
Please also note these General Information Points   また、以下の一般的な情報にもご留意いただきたい。
There may be additional rules and requirements applicable to cloud computing in specific sectors which should also be considered (for example, in the financial, health, telecom, industry, energy or public sector, to name a few).   特定のセクター(例えば、金融、医療、通信、産業、エネルギー、公共セクターなど)において、クラウド・コンピューティングに適用される追加の規則や要件が存在する可能性もある。
Throughout the relation and legal lifecycle between Customer and Provider, keep in mind the obligation of good faith on the parties, as mentioned explicitly in Article 27 of the Data Act.   顧客とプロバイダ間の関係および法的ライフサイクルを通じて、データ法第27条に明記されている当事者の誠実義務に留意すること。
Please also note that the Data Act provides for exceptions to certain obligations related to the SCCs Switching & Exit. Reference is made to Article 31 Data Act. Further explanations can also be found in the SCCs Switching & Exit.  また、データ法は、SCCsのスイッチング&エグジットに関連する特定の義務に対する例外をプロバイダとして規定していることに留意されたい。データ法第31条を参照のこと。さらに詳しい説明はSCCs Switching & Exitにも記載されている。

 

 

 

| | Comments (0)

2025.04.02

文部科学省 教育情報セキュリティポリシーに関するガイドライン(令和7年3月) (2025.03.25)

こんにちは、丸山満彦です。

文部科学省が教育情報セキュリティポリシーに関するガイドライン教育情報セキュリティポリシーハンドブックを改訂し、公表していますね...

文部科学省の教育情報セキュリティポリシーに関するガイドラインは、総務省の自治体ガイドラインを参照している部分がありますね..(総務省のガイドラインも改訂されたのでリンクが切れてしまっていますが...)

情報資産の分類・仕分け・管理方法の見直し及び次世代校務DX環境への移行を進める上で必要となるセキュリティ対策に関する記載の見直しが必要ということでの変更ということのようです...

 

次世代校務DX環境...

次世代校務DX(クラウド上での校務実施を前提とし、ロケーションフリーやデータ利活用・データ連携を通じて、学校の働き方改革・教育活動の高度化・教育現場におけるレジリエンス確保の実現に資する新しい校務の在り方)を実現するために整備が必要となる環境

ガイドラインはこちらも200ページ越えですね...

 

 過去の改訂履歴と理由...(表紙の履歴では令和元年の改訂が抜けていますね...)

・2017年;平成29年10月 策定;教育現場には、児童生徒や保護者の存在等、地方公共団体の他の行政事務とは異なる特徴があることから、これらを考慮した情報セキュリティ対策を講じる必要

・2019年;令和01年12月 改訂;GIGAスクール構想の始動時に対応するため改訂

・2021年;令和03年05月 改訂;新たに必要なセキュリティ対策やクラウドサービスの活用を前提としたネットワーク構成等の課題に対応するため改訂

・2022年;令和04年03月 一部改訂;アクセス制御による対策の詳細な技術的対策の追記と、「ネットワーク分離による対策」、「アクセス制御による対策」を明確に記述するため改訂

・2024年;令和06年01月 改訂;次世代の校務DX環境の整備を見据えた情報セキュリティの考え方の提示、関連法令・指針の改訂・改正に伴う対応、各自治体における教育情報セキュリティポリシーの策定推進に向けた読みやすさ向上のため改訂

・2025年;令和07年03月 改訂;情報資産の分類・仕分け・管理方法の見直し及び次世代校務DX環境への移行を進める上で必要となるセキュリティ対策に関する記載の見直し

教育情報セキュリティポリシーに関するガイドライン(令和7年3月)

・[PDF] 教育情報セキュリティポリシーに関するガイドライン(令和7年3月) 

20250402-35639

 

・[PDF] 【見え消し版】教育情報セキュリティポリシーに関するガイドライン(令和7年3月)

20250402-35737

 

・[PDF] 教育情報セキュリティポリシーに関するガイドライン(令和7年3月)改訂説明資料 

20250402-40030

 

教育情報セキュリティポリシーハンドブック(令和7年3月)

・[PDF] 教育情報セキュリティポリシーハンドブック(令和7年3月) 

20250402-35829

 

 


 

 まるちゃんの情報セキュリティ気まぐれ日記

・2022.03.05 文部科学省 教育情報セキュリティポリシーに関するガイドライン

・2021.05.31 文部科学省 「教育情報セキュリティポリシーに関するガイドライン」公表について

 

総務省のガイドライン...

● まるちゃんの情報セキュリティ気まぐれ日記

・2025.04.01 総務省 地方公共団体における情報セキュリティポリシーに関するガイドライン(令和7年3月版)

・2020.12.29 総務省 「地方公共団体における情報セキュリティポリシーに関するガイドライン」と「地方公共団体における情報セキュリティ監査に関するガイドライン」の公表及び意見募集の結果

・2020.12.13 総務省 意見募集「地方公共団体における情報セキュリティポリシーに関するガイドライン」(改定案)、「地方公共団体における情報セキュリティ監査に関するガイドライン」(改定案)

・2020.05.23 総務省 「自治体情報セキュリティ対策の見直しについて」の公表

 

ちょっと遡って...

・2010.11.16 総務省 確定 地方公共団体における情報セキュリティポリシーに関するガイドラインと地方公共団体における情報セキュリティ監査に関するガイドライン

・2010.09.17 総務省 パブコメ 地方公共団体における情報セキュリティポリシーに関するガイドライン(案)と地方公共団体における情報セキュリティ監査に関するガイドライン(案)

・2010.08.06 総務省 電子自治体 情報セキュリティ対策の推進

・2007.07.09 総務省 確定 「地方公共団体における情報セキュリティ監査に関するガイドライン」

・2007.06.09 総務省 パブコメ 「地方公共団体における情報セキュリティ監査に関するガイドライン」

・2007.04.02 総務省 自治体ISAC(仮称)実証実験の実施結果

・2006.09.30 総務省 地方公共団体における情報セキュリティポリシーに関するガイドラインを公表

・2006.08.21 総務省 パブコメ 「地方公共団体における情報セキュリティポリシーに関するガイドライン」(案)

・2006.06.01 地方公共団体セキュリティ対策支援フォーラム 「情報セキュリティ監査実施状況および推進課題に関する検討」報告書

・2006.04.06 総務省 「地方公共団体の情報セキュリティレベルの評価に係る制度の在り方に関する調査研究報告書」の公表

・2005.07.19 総務省 平成17年度電子自治体関連施策 セキュリティ認定制度

・2005.07.17 「地方公共団体における情報セキュリティ内部アセスメント(監査)の進め方」

 

 

| | Comments (0)

2025.04.01

経済産業省 営業秘密管理指針の改訂(2025.03.31)

こんにちは、丸山満彦です。

経済産業省が、営業秘密管理指針を改訂していますね...営業秘密管理指針は、「不正競争防止法」による保護を受けるために必要となる最低限の水準の対策と経済産業省が考えるところを示すものですね...

経済産業省 - 営業秘密~営業秘密を守り活用する~

・2025.03.31 [PDF] 営業秘密管理指針(令和7年3月改訂版)

20250401-61542

目次...

はじめに(本指針の性格)
1.総説
2.秘密管理性について

(1) 秘密管理性要件の趣旨
(2) 必要な秘密管理措置の程度
(3) 秘密管理措置の具体例
・紙媒体の場合
・電子媒体の場合
・物件に営業秘密が化体している場合
・媒体が利用されない場合
・複数の媒体で同一の営業秘密を管理する場合

(4) 営業秘密を企業内外で共有する場合の秘密管理性の考え方
・社内の複数箇所で同じ情報を保有しているケース
・複数の法人間で同一の情報を保有しているケース

3.有用性の考え方
4.非公知性の考え方
おわりに


改訂の経緯...

  • 2003年;平成15年01月30日;策定;「企業が営業秘密に関する管理強化のための戦略的なプログラムを策定できるよう、参考となるべき指針」として策定
  • 2005年;平成17年10月12日;改訂;
  • 2010年;平成22年04月09日;改訂;
  • 2011年;平成23年12月01日;改訂;
  • 2013年;平成25年08月16日;改訂;
  • 2015年;平成27年01月28日;改訂;一部の裁判例等において秘密管理性の認定が厳しいとの指摘や認定の予見可能性を高めるべき、事業者にとってより分かりやすい記載とするよう改訂
  • 2019年;平成31年01月23日;改訂;ビッグデータ、AI の活用が推進する第四次産業革命を背景として情報活用形態が多様化する状況を踏まえ改訂
  • 2025年;令和07年03月31日;改訂;
    • テレワーク勤務など、企業の施設外における労働の機会が増えており、これに伴い、自宅等において営業秘密に触れる機会が増えている
    • 派遣労働者が、営業秘密に接する機会も増えている
    • 兼業・副業の動きも見られ、兼業先・副業先の営業秘密に接する機会が生じている
    • 一情報管理の手法については、クラウド技術・環境を前提とした管理が進むなど、企業における情報管理のあり方のが変化している

ということのようです...

 

関連指針等...

・2024.02 [PDF] 限定提供データに関する指針

20250401-63537

 

・2024.02 [PDF] 秘密情報の保護ハンドブック

290ページあります...

20250401-63725

 

 


 

まるちゃんの情報セキュリティ気まぐれ日記

・2022.05.18 経済産業省 産業構造審議会 知的財産分科会 不正競争防止小委員会 中間報告書、秘密情報の保護ハンドブック等

 

・2011.12.02 経済産業省 「営業秘密管理指針(改訂版)」

・2010.04.10 経済産業省 確定 営業秘密管理指針

・2010.02.17 経済産業省 パブコメ 「営業秘密管理指針の再改訂(案)」

・2005.10.13 経済産業省 営業秘密管理指針の改訂

・2005.09.15 経済産業省 バプコメ 営業秘密管理指針改訂(案)

 

| | Comments (0)

総務省 地方公共団体における情報セキュリティポリシーに関するガイドライン(令和7年3月版)

こんにちは、丸山満彦です。

総務省が地方公共団体における情報セキュリティポリシーに関するガイドライン(令和7年3月版)を公表しています...

意見募集が166件もあったので、内容を見たのですが、言わんとしている意図はわかるのですが、伝え方はどうかなというものも多くあるように思いました。

ただ、ポリシーに関するガイドラインが413ページ、監査ガイドラインが189ページで、流石にこれを読んで理解するのは、対象が市政都市とするならばしんどいだろうなとは思いました。

また、1週間のパブコメ期間というのも短いという気がしました。

改訂履歴はつぎのような感じですかね...

地方公共団体における情報セキュリティポリシーに関するガイドライン

  • 2001年;平成13年03月;策定
  • 2006年;平成18年09月;全部改定;政府の情報セキュリティ政策会議は「第 1 次情報セキュリティ基本計画」を踏まえて改定
  • 2010年;平成22年11月;一部改定;「第 2 次情報セキュリティ基本計画」等を踏まえて改定
  • 2015年;平成27年03月;一部改定;マイナンバー法、サイバーセキュリティ基本法を踏まえて改定
  • 2018年;平成30年03月;全部改定;平成27年の年金機構情報漏えい事案、三層対策
  • 2020年;令和02年12月;一部改定;三層対策のパターンを増やすなど利便性とセキュリティのバランスの調整
  • 2022年;令和04年03月;一部改定;政府統一基準の改訂、DX等を踏まえた改定
  • 2023年;令和05年03月;一部改定;ガバメントクラウドの導入等を見据えて改定
  • 2024年;令和06年10月;一部改定;Web会議利用、政府統一基準の改定を踏まえた改定
  • 2025年;令和07年03月;一部改定;テレワーク、マイナンバー利用事務系への無線LAN接続等を踏まえた改定

地方公共団体における情報セキュリティ監査に関するガイドライン

  • 2003年;平成15年12月;策定
  • 2007年;平成19年07月;全部改定;政府の情報セキュリティ政策会議は「第 1 次情報セキュリティ基本計画」を踏まえて改定
  • 2010年;平成22年11月;一部改定;「第 2 次情報セキュリティ基本計画」等を踏まえて改定
  • 2015年;平成27年03月;一部改定;マイナンバー法、サイバーセキュリティ基本法を踏まえて改定
  • 2018年;平成30年09月;一部改定;政府統一基準、自治体情報セキュリティ対策検討チーム報告等を踏まえて改訂
  • 2020年;令和02年12月;一部改訂;三層対策のパターンを増やすなどの対策の変化に応じて改訂
  • 2022年;令和04年03月;一部改訂;政府統一基準の改訂、DX等を踏まえた改定

 初版からの積み重ねということもあり、どんどんページ数が増えていっている感じですよね...

読者層を首長等、管理職、システム担当者、全従業員のように分けて記載できれば、ぐっと楽になるような気がします...(政府統一基準の初版では主語を定義して、誰が実施すべきことかわかるようにしたんですけどね...)

 

● 総務省

・2025.03.28 「地方公共団体における情報セキュリティポリシーに関するガイドライン」等の意見募集の結果及び改定版の公表

 

・・[PDF] 別紙1 「地方公共団体における情報セキュリティポリシーに関するガイドライン」等の改定について

20250401-12052

・・[PDF] 別紙2 「地方公共団体における情報セキュリティポリシーに関するガイドライン」

20250401-12012

・・[PDF] 別紙3 「地方公共団体における情報セキュリティ監査に関するガイドライン」

20250401-12017

・・[PDF] 別紙4 「地方公共団体における情報セキュリティポリシーに関するガイドライン」 別紙 マイナンバー利用事務系に係る画面転送の方式について

20250401-12059

・・[PDF] 別紙5 「地方公共団体における情報セキュリティ監査に関するガイドライン」別紙 マイナンバー利用事務系に係る画面転送の方式について 参考資料 「マイナンバー利用事務系に係る画面転送の方式」監査項目

20250401-12105

・・[PDF]
別紙6 「地方公共団体における情報セキュリティポリシーに関するガイドライン」(改定案)等に対する意見募集結果について

 

 

 

 

 

 

 

 


 

● まるちゃんの情報セキュリティ気まぐれ日記

・2021.09.14 地⽅⾃治体によるガバメントクラウドの活⽤について(案)2021.08現在

・2021.05.31 文部科学省 「教育情報セキュリティポリシーに関するガイドライン」公表について

・2021.05.13 参議院 デジタル社会形成基本法案等を議決 デジタル庁設置、個人情報保護法改正、地方自治体システム標準化等。。。

・2020.12.31 自民党デジタル社会推進本部 デジタル庁創設に向けた中間提言 at 2020.12.22 (小林史明議員公式サイト)

・2020.12.29 総務省 「地方公共団体における情報セキュリティポリシーに関するガイドライン」と「地方公共団体における情報セキュリティ監査に関するガイドライン」の公表及び意見募集の結果

・2020.12.25 官邸 「デジタル社会の実現に向けた改革の基本方針」と「デジタル・ガバメント実行計画」が閣議決定

・2020.12.13 総務省 意見募集「地方公共団体における情報セキュリティポリシーに関するガイドライン」(改定案)、「地方公共団体における情報セキュリティ監査に関するガイドライン」(改定案)

・2020.11.19 自民党デジタル社会推進本部がデジタル庁についての第一次提言を平井卓也デジタル改革担当相に手交

・2020.09.05 J-LIS (予告)「自治体テレワーク推進実証実験」の公募について

・2020.05.23 総務省 「自治体情報セキュリティ対策の見直しについて」の公表

 

ちょっと遡って...

・2012.07.22 総務省 ASP・SaaS・クラウドの普及拡大に向けたガイドの公表

・2010.11.16 総務省 確定 地方公共団体における情報セキュリティポリシーに関するガイドラインと地方公共団体における情報セキュリティ監査に関するガイドライン

・2010.12.12 総務省 自治体クラウド推進本部 有識者懇談会(第3回)

・2010.09.17 総務省 パブコメ 地方公共団体における情報セキュリティポリシーに関するガイドライン(案)と地方公共団体における情報セキュリティ監査に関するガイドライン(案)

・2010.08.06 総務省 電子自治体 情報セキュリティ対策の推進

・2010.08.05 総務省 自治体クラウド推進本部

・2010.05.01 内閣府 地方公共団体の業務継続ガイドライン

・2010.04.02 総務省 確定 地方公共団体におけるASP・SaaS導入活用ガイドライン

・2010.03.12 「サイバー攻撃に無防備、193自治体」だそうです。。。

・2010.03.09 総務省 自治体クラウドポータルサイトの開設

・2010.02.20 総務省 パブコメ 「地方公共団体におけるASP・SaaS導入活用ガイドライン(案)」

・2010.01.25 長崎県がクラウド事業者として市町村にサービスを提供するの件

・2010.01.18 日本経団連 「電子行政推進シンポジウム」(2009.12.08)開催の様子

・2009.03.28 総務省 電子自治体の推進に関する懇談会(セキュリティワーキング グループ)検討結果

・2008.08.23 総務省 確定 「地方公共団体におけるICT部門の業務継続計画(BCP)策定に関するガイドライン」

・2008.06.28 総務省 パブコメ 「地方公共団体におけるICT部門の業務継続計画(BCP)策定に関するガイドライン」(案)

・2008.04.26 総務省 地方自治情報管理概要

・2007.07.14 総務省 「地方公共団体におけるITガバナンスの強化ガイド」を公表

・2007.07.09 総務省 確定 「地方公共団体における情報セキュリティ監査に関するガイドライン」

・2007.06.09 総務省 パブコメ 「地方公共団体における情報セキュリティ監査に関するガイドライン」

・2007.04.02 総務省 自治体ISAC(仮称)実証実験の実施結果

・2006.09.30 総務省 地方公共団体における情報セキュリティポリシーに関するガイドラインを公表

・2006.08.21 総務省 パブコメ 「地方公共団体における情報セキュリティポリシーに関するガイドライン」(案)

・2006.06.01 地方公共団体セキュリティ対策支援フォーラム 「情報セキュリティ監査実施状況および推進課題に関する検討」報告書

・2006.04.06 総務省 「地方公共団体の情報セキュリティレベルの評価に係る制度の在り方に関する調査研究報告書」の公表

・2006.01.24 総務省 住民基本台帳ネットワークシステム及びそれに接続している既設ネットワークに関する調査票による点検状況

・2005.07.19 総務省 平成17年度電子自治体関連施策 セキュリティ認定制度

・2005.07.17 「地方公共団体における情報セキュリティ内部アセスメント(監査)の進め方」

・2005.02.15 住民基本台帳ネットワークシステム 政府と国民の信頼がポイントではないのだろうか?

 

| | Comments (0)

2025.03.26

米国 NIST SP 800-228(初期公開ドラフト)クラウドネイティブシステムのAPI防御ガイドライ

こんにちは、丸山満彦です。

NISTがSP 800-228(初期公開ドラフト)クラウドネイティブシステムのAPI保護ガイドライを公表し、意見募集をしていますね...

組織のシステムがクラウドに移行に、各システムの連携がAPIを通じて行われるようになると、APIをいかに守るかということは開発者にとっても、利用者にとっても重要となりますよね...

 

NIST - ITL

・2025.03.25 NIST SP 800-228 (Initial Public Draft) Guidelines for API Protection for Cloud-Native Systems

 

NIST SP 800-228 (Initial Public Draft) Guidelines for API Protection for Cloud-Native Systems NIST SP 800-228(初期公開ドラフト)クラウドネイティブシステムのAPI保護ガイドライ
Announcement 発表
Modern enterprise IT systems rely on a family of application programming interfaces (APIs) for integration to support organizational business processes. Hence, a secure development and deployment of APIs is critical for overall enterprise security. This, in turn, requires the identification of risk factors or vulnerabilities in various phases of the API life cycle and the development of controls or protection measures to prevent their exploits. 現代のエンタープライズITシステムは、組織のビジネスプロセスをサポートする統合のために、アプリケーション・プログラミング・インターフェース(API)ファミリーに依存している。したがって、APIの安全な開発と展開は、エンタープライズ全体のセキュリティにとって極めて重要である。そのためには、APIのライフサイクルの様々な段階におけるリスク要因や脆弱性を特定し、その悪用を防ぐための管理策や防御策を開発する必要がある。
This document addresses the following aspects for achieving that goal: この文書では、その目標を達成するために、次の側面を取り上げる:
a. The identification and analysis of risk factors or vulnerabilities introduced during various activities of API development and runtime, a. API の開発と実行時の様々な活動中に導入されるリスク要因や脆弱性の特定と分析、
b. Recommended basic and advanced controls and protection measures during the pre-runtime and runtime stages of APIs, and b. API の実行前と実行時の段階における、推奨される基本的で高度な制御と保護対策、
c. An analysis of the advantages and disadvantages of various implementation options (i.e., patterns) for those controls to enable security practitioners to adopt an incremental, risk-based approach to securing their APIs. c. セキュリティ担当者が、API の安全性を確保するために、段階的でリスクに基づいたアプローチを採用できるようにするための、 これらの制御の様々な実装オプション(すなわち、パターン)の利点と欠点の分析。
Abstract 概要
Modern enterprise IT systems rely on a family of application programming interfaces (APIs) for integration to support organizational business processes. Hence, a secure deployment of APIs is critical for overall enterprise security. This, in turn, requires the identification of risk factors or vulnerabilities in various phases of the API life cycle and the development of controls or protection measures. This document addresses the following aspects of achieving that goal: (a) the identification and analysis of risk factors or vulnerabilities during various activities of API development and runtime, (b) recommended basic and advanced controls and protection measures during pre-runtime and runtime stages of APIs, and (c) an analysis of the advantages and disadvantages of various implementation options for those controls to enable security practitioners to adopt an incremental, risk-based approach to securing their APIs. 現代のエンタープライズITシステムは、組織のビジネスプロセスをサポートするために、アプリケーション・プログラミング・インターフェース(API)ファミリーの統合に依存している。したがって、APIの安全な展開は、エンタープライズ全体のセキュリティにとって極めて重要である。そのためには、API のライフサイクルの様々な段階におけるリスク要因や脆弱性を特定し、管理策や保護策を開発する必要がある。この文書では、その目標を達成するために、以下の側面を取り上げる:(a) API の開発及び実行時の様々な活動におけるリスク要因又は脆弱性の特定と分析、(b) API の実行前及び実行時の段階における、推奨される基本的及び高度な管理及び保護対策、(c) セキュリティ実務者が API の安全性を確保するために、段階的でリスクに基づいたアプローチを採用できるようにするための、これらの管理に関する様々な実装オプションの利点と欠点の分析。

 

・[PDF] NIST.SP.800-228.ipd

20250326-52859

 

目次...

Executive Summary エグゼクティブサマリー
1. Introduction 1. 序文
1.1. Building Blocks and Structures 1.1. ビルディングブロックと構造
1.2. Zero Trust and APis: The Vanishing Perimeter 1.2. ゼロトラストとAPI: 
1.3. API Life Cycle 1.3. APIのライフサイクル
1.4. Document Goals… 1.4. 文書の目標...
1.5. Relationship to Other NIST Documents 1.5. 他の NIST 文書との関係
1.6. Document Structure 1.6. 文書構造
2. API Risks — Vulnerabilities and Exploits 2. API リスク - 脆弱性とエクスプロイト
2.1. Lack of Visibility of APis in the Enterprise Inventory 2.1. エンタープライズインベントリにおける API の可視性の欠如
2.2. Missing, Incorrect, or Insufficient Authorization 2.2. 認可の欠落、不正確、不十分
2.3. Broken Authentication 2.3. 破られた認証
2.4. Unrestricted Resource Consumption 2.4. 無制限のリソース消費
2.4.1. Unrestricted Compute Resource Consumption 2.4.1. 無制限の計算リソース消費
2.4.2. Unrestricted Physical Resource Consumption  2.4.2. 制限のない物理リソースの消費
2.5. Leaking Sensitive Information to Unauthorized Callers 2.5. 権限のない発信者への機密情報の漏洩
2.6. Insufficient Verification of Input Data 2.6. 入力データの不十分な検証
2.6.1. Input Validation 2.6.1. 入力妥当性確認
2.6.2. Malicious Input Protection 2.6.2. 悪意のある入力保護
2.7. Credential Canonicalization- Preparatory Step for Controls 2.7. クレデンシャルの正規化 - コントロールの準備ステップ
2.7.1. Gateways Straddle Boundaries 2.7.1. ゲートウェイは境界をまたぐ
2.7.2. Requests With a Service Identity but No User ldentity 2.7.2. サービスIDを持つがユーザ IDを持たないリクエスト
2.7.3. Requests With a User Identity But No Service Identity 2.7.3. ユーザーIDを持つがサービスIDを持たないリクエスト
2.7.4. Requests With Both User and Service Identities 2.7.4. ユーザーIDとサービスIDの両方を持つリクエスト
2.7.5. Reaching Out to Other Systems  2.7.5. 他のシステムへのリーチアウト
2.7.6. Mitigating the Confused Deputy 2.7.6. 混乱した代理の緩和
2.7.7. Identity Canonicalization 2.7.7. アイデンティティの正規化
3. Recommended Controls for APls 3. APls
3.1. Pre-Runtime Protections 3.1. 実行前の防御
3.1.1. Basic Pre-Runtime Protections 3.1.1. 基本的なプレランタイム防御
3.1.2. Advanced Pre-Runtime Protections 3.1.2. 高度なプレランタイム防御
3.2. Runtime Protections 3.2. ランタイム防御
3.2.1. Basic Runtime Protections 3.2.1. 基本的なランタイム防御
3.2.2. Advanced Runtime Protections 3.2.2. 高度なランタイム防御
4. Implementation Patterns and Trade-Offs for API Protections 4. API防御の実装パターンとトレードオフ
4.1. Centralized API Gateway 4.1. 集中型APIゲートウェイ
4.2. Hybrid Deployments 4.2. ハイブリッド展開
4.3. Decentralized Gateways 4.3. 非中央集権化ゲートウェイ
4.4. Related Technologies 4.4. 関連技術
4.4.1. Web Application Firewalls 4.4.1. ウェブアプリケーションファイアウォール
4.4.2. Bot Detection 4.4.2. ボット検知
4.4.3. Distributed Denial of Service (DDoS) Mitigation 4.4.3. 分散型サービス拒否(DDoS)緩和
4.4.4. API Endpoint Protection, 4.4.4. API エンドポイント防御、
4.4.5. Web Application and API Protection (WAAP) 4.4.5. ウェブアプリケーション・API防御(WAAP)
4.5. Summary of Implementation Patterns 4.5. 実装パターンのまとめ
5. Conclusions and Summary 5. 結論とまとめ
References 附属書
Appendix A. API Classification Taxonomy  附属書A. APIの分類分類法
A.1. API Classification Based on Degree of Exposure A.1. エクスポージャーの程度に基づくAPIの分類
A.2. API Classification Based on Communication Patterns A.2. コミュニケーションパターンに基づくAPI分類
A.3. API Classification Based on Architectural Style or Pattern (API Types) A.3. アーキテクチャスタイルまたはパターン(APIタイプ)に基づくAPI分類
Appendix B. DevSecOps Phase and Associated Class of API Controls 附属書B. DevSecOpsのフェーズとAPIコントロールの関連クラス

 

 

エグゼクティブサマリーと序文...

Executive Summary  エグゼクティブサマリー
Application programming interfaces (APIs) provide the means to integrate and communicate with the modern enterprise IT application systems that support business processes. However, a lack of due diligence can introduce vulnerabilities and risk factors that exploit the connectivity and accessibility features of APIs. If these vulnerabilities are not identified, analyzed and addressed through control measures, attack vectors could threaten the security posture of the application systems spanned by these APIs. A systematic and effective means of identifying and addressing these vulnerabilities is only possible by treating the development and deployment of APIs as an iterative life cycle using paradigms like DevSecOps. アプリケーションプログラミングインタフェース(API)は、ビジネスプロセスをサポートする最新のエンタープライズITアプリケーションシステムと統合し、コミュニ ケーションするための手段を提供する。しかし、デューディリジェンスの欠如は、APIの接続性とアクセシビリティ機能を悪用する脆弱性とリスク要因を導入する可能性がある。これらの脆弱性が特定され、分析され、管理策を通じて対処されない場合、攻撃ベクトルは、これらのAPIによってスパンされるアプリケーションシステムのセキュリティ体制を脅かす可能性がある。これらの脆弱性を特定し、対処する体系的で効果的な手段は、DevSecOps のようなパラダイムを使用して、 API の開発と展開を反復的なライフサイクルとして扱うことによってのみ可能である。 
This document provides guidance and recommendations on controls and protection measures for secure API deployments in the enterprise. In addition, an analysis of the advantages and disadvantages of various implementation options (called patterns) for those controls enable security practitioners to choose the most effective option for their IT ecosystem.  本文書は、エンタープライズにおける安全なAPI展開のためのコントロールと保護対策に関するガイダンスと推奨を提供する。加えて、これらの管理策の様々な実装オプション(パターンと呼ばれる)の長所と短所を分析することで、セキュリティ担当者が、ITエコシステムに最も効果的なオプションを選択できるようにする。
Developing these controls and analyzing their implementation options should be guided by several overarching principles: このような管理策の策定と実装オプションの分析は、いくつかの包括的な原則によって導 かれなければならない: 
The guidance for controls should cover all APIs, regardless of whether they are exposed to customers/partners or used internally within the enterprise.  管理策のガイダンスは、API が顧客/パートナーに公開されているか、エンタープライズ内部で使用されているかに関係なく、すべての API を対象とすべきである。
With the vanishing of perimeters in modern enterprise IT applications, all controls should incorporate the concept of zero trust.  現代のエンタープライズITアプリケーションでは境界がなくなりつつあるため、全ての管理はゼロトラストの概念を取り入れるべきである。
The controls should span the entire API life cycle and be classified into (a) pre-runtime protections and (b) runtime protections that are then subdivided into basic and advanced protections to enable incremental risk-based adoption.  コントロールはAPIのライフサイクル全体に及び、(a)実行前の防御と(b)実行時の防御に分類され、さらにリスクベースの段階的な導入を可能にするために、基本的な防御と高度な防御に細分化されるべきである。
1. Introduction 1. 序文
Application programming interfaces (APIs) represent an abstraction of the underlying implementation of a digital enterprise. Given the spatial (e.g., on-premises, multiple clouds) and logical (e.g., microservices) nature of current enterprise applications, APIs are needed to integrate and establish communication pathways between internal and third-party services and applications. Informally, APIs are the lingua franca of modern IT systems: they describe what actions users are allowed to take. They are also used in every type of application, including server-based monolithic, microservices-based, browser-based client, and IoT.  アプリケーション・プログラミング・インターフェース(API)は、デジタル・エ ンタープライズの基本的な実装を抽象化した代表者である。現在のエンタープライズ・アプリケーションの空間的(例:オンプレミス、複数のクラウド)かつ論理的(例:マイクロサービス)な性質を考慮すると、APIは内部とサードパーティ・サービスおよびアプリケーション間のコミュニケーション経路を統合し確立するために必要である。非公式に言えば、APIは現代のITシステムの共通語であり、ユーザーがどのようなアクションを取ることができるかを記述している。また、サーバーベースのモノリシック、マイクロサービスベース、ブラウザベースのクライアント、IoTなど、あらゆるタイプのアプリケーションで使用されている。
1.1. Building Blocks and Structures 1.1. ビルディング・ブロックと構造
An Application Programming Interface (API) defines how any two pieces of software communicate – they are ubiquitous in software. An API is a collection of commands or endpoints that operate on data or objects via some protocol. Network-based APIs are APIs built to be consumed by remote applications over the network. Because they’re exposed and consumed over the network, they present a unique set of challenges. The growth of (micro-) service-oriented architectures, coupled with Software-as-a-Service (SaaS) becoming commonplace – which are nearly always delivered via APIs – has resulted in an explosion in network-based APIs across organizations. This document focuses on controls for network-based APIs.  アプリケーション・プログラミング・インターフェース(API)は、2つのソフトウェアがどのようにコミュニケーションするかを定義する。APIは、何らかのプロトコルを介してデータやオブジェクトを操作するコマンドやエンドポイントの集合体である。ネットワークベースのAPIは、ネットワークを介してリモート・アプリケーションから消費されるように作られたAPIだ。ネットワーク上で公開され、消費されるため、APIには独特の課題がある。(マイクロ)サービス指向アーキテクチャの成長と、SaaS(Software-as-a-Service)の一般化(これらはほぼ常にAPI経由で提供される)が相まって、組織全体でネットワークベースのAPIが爆発的に増加した。この文書では、ネットワークベースの API のコントロールに焦点を当てる。
Before we can discuss API controls, we need a common understanding and language for the building blocks, and how they relate to each other. The taxonomy is: an API is composed of a set of API Endpoints; API Endpoints are implemented by Services; at runtime, Requests to a specific API Endpoint are served by Service Instances. An API Gateway hosts many APIs and is responsible for mapping each Request to its target API Endpoint, applying policy for that Endpoint (e.g. authentication and rate limiting), then routing that Request to a Service Instance which implements that API Endpoint.  APIのコントロールについて議論する前に、構成要素とそれらがどのように互いに関連しているかについての共通の理解と言語が必要である。APIはAPIエンドポイントのセットで構成され、APIエンドポイントはサービスによって実装され、実行時に特定のAPIエンドポイントへのリクエストはサービスインスタンスによって処理される。APIゲートウェイは多くのAPIをホストし、各リクエストをターゲットのAPIエンドポイントにマッピングし、そのエンドポイントのポリシー(認証やレート制限など)を適用し、そのAPIエンドポイントを実装するサービスインスタンスにそのリクエストをルーティングする。
20250326-53734
Fig. 1. API, API Endpoint, Service and Service Instance  Fig. API、APIエンドポイント、サービス、サービスインスタンス
Traditionally, we think of network-based APIs as being customer-oriented, partner-oriented, or internal – often called “third-party”, “second-party”, and “first-party” APIs, respectively. Second- and third-party APIs are typically exposed to callers outside of the organization via an API gateway. First-party APIs can be exposed to callers inside of the organization on the same API gateway, but they are also often consumed directly by internal callers without traversing a dedicated API serving stack.  伝統的に、我々はネットワークベースのAPIを顧客指向、パートナー指向、内部指向、それぞれ「サードパーティ」、「セカンドパーティ」、「ファーストパーティ」APIと呼ばれることが多いと考える。セカンドパーティとサードパーティのAPIは通常、APIゲートウェイを介して組織外の呼び出し元に公開される。ファーストパーティAPIは同じAPIゲートウェイ上で組織内部の呼び出し元に公開されることもあるが、専用のAPIサービングスタックを経由せずに内部の呼び出し元によって直接消費されることも多い。
An API is a set of API Endpoints, and a Service implements a set of API Endpoints – so every Service implements some API. We call these Service APIs. Most first-party API integrations happen via the Service API, i.e. they map to a single service. On the other hand, APIs hosted by the API gateway typically have Endpoints that map to many different Services. This is especially common for second- and third-party APIs. We call these Facade APIs, because they present a single facade to an outside caller over (potentially many) different Service APIs. Finally, it’s common that multiple Services are grouped together into an Application – typically along organizational lines (often an Application maps to a team). Schematic diagrams of a Service API, Façade API and an Application (Monolithic) API are given below:  APIはAPIエンドポイントの集合であり、サービスはAPIエンドポイントの集合を実装する。私たちはこれらをサービスAPIと呼んでいる。ほとんどのファーストパーティAPI統合はサービスAPIを介して行われる。一方、APIゲートウェイによってホストされるAPIは通常、多くの異なるサービスにマッピングされるエンドポイントを持っている。これは特にセカンドパーティやサードパーティのAPIによく見られる。このようなAPIをファサードAPIと呼ぶ。なぜなら、ファサードAPIは、(潜在的に多くの)異なるサービスAPI上の単一のファサードを外部の呼び出し元に提示するからである。最後に、複数のサービスがアプリケーションにグループ化されるのはよくあることで、通常は組織的な線に沿っている(多くの場合、アプリケーションはチームにマッピングされる)。サービスAPI、ファサードAPI、アプリケーション(モノリシック)APIの概略図を以下に示す: 
20250326-53845
Fig. 2. (Top to Bottom) Service API, Façade API, Service and Application (Monolithic)  図2(上から下へ)サービスAPI、ファサードAPI、サービス、アプリケーション(モノリシック)
Less formally: we can think of the APIs we expose outside the organization as a facade over a set of Services. Those Services implement internal APIs (Service APIs). Services in the organization communicate with each other via those internal APIs – sometimes directly, and sometimes via an API gateway. The API Gateway is responsible for some policies, like authentication and rate limiting, as well as being responsible for mapping the facade APIs for external clients to internal APIs. Then, to get a handle on things organizationally, we often group related Services into a bucket called an Application.  あまり形式的ではないが、組織の外部に公開するAPIは、一連のサービスを覆うファサードと考えることができる。これらのサービスは内部API(サービスAPI)を実装する。組織内のサービスは、これらの内部APIを介して、時には直接、時にはAPIゲートウェイを介して相互にコミュニケーションする。APIゲートウェイは、認証やレート制限のようなポリシーを担当し、外部クライアントのためのファサードAPIを内部APIにマッピングする役割も担っている。それから、組織的に物事を把握するために、関連するサービスをアプリケーションと呼ばれるバケットにグループ化することが多い。
While we tend to think of APIs in the context of exposing functionality to clients or partners, APIs don’t exist solely at the edge of our infrastructure. Any time systems communicate, there’s some API involved. Even if that API is something like CSV over FTP. The examples in SP focus primarily on “modern” APIs exposed via mechanisms like HTTP/REST, gRPC, or SOAP, but we believe the principals in this SP are universal and should be applied to all APIs.  APIというと、クライアントやパートナーに機能を公開するという文脈で考えがちだが、APIはインフラの端だけに存在するわけではない。システムがコミュニケーションするときには、必ずAPIが関係している。たとえそのAPIがFTP経由のCSVのようなものであってもだ。SPの例は、主にHTTP/REST、gRPC、SOAPのようなメカニズムで公開される 「最新の 」APIに焦点を当てているが、このSPの原則は普遍的であり、すべてのAPIに適用されるべきだと考えている。
1.2. Zero Trust and APIs: The Vanishing Perimeter  1.2. ゼロトラストとAPI 消滅する境界
APIs are built out of services that communicate with each other via APIs, similar to how the internet is a “network of networks.” One of the most important implications of zero trust is that there is no meaningful distinction between an “internal” and “external” caller because the perimeter is the service instance itself. Rather, all callers are trusted if they are authorized to be trusted. This contrasts with traditional approaches to API security in which the only “APIs” are those exposed to “external” callers, and API-oriented controls are only enforced at the perimeter, typically via an API gateway.  APIは、インターネットが 「ネットワークのネットワーク 」であるのと同様に、APIを介して相互にコミュニケーショ ンするサービスから構築される。ゼロトラストの最も重要な意味の1つは、境界がサービスインスタンスそのものであるため、「内部」と「外部」の呼び出し者の間に意味のある区別がないということである。むしろ、すべての呼び出し元は、信頼されることが認可されていれば信頼される。これは、「API」だけが「外部」呼び出し元に公開され、API指向の制御が境界でのみ、典型的にはAPIゲートウェイを介して実施される、APIセキュリティへの従来のアプローチとは対照的である。
NIST Special Publication (SP) 800-207A [6] discusses zero trust at runtime and the principle of shrinking the perimeter to the service instance using the five runtime controls of identity-based segmentation:  NIST 特別刊行物(SP)800-207A [6]は、実行時のゼロトラストと、ID ベースのセグメンテーションの 5 つの実行時コントロールを使用して、境界をサービスインスタンスに縮小する原則について論じている: 
1. Encryption in transit — To ensure message authenticity and prevent eavesdropping, thus preserving confidentiality  1. 転送中の暗号化 - メッセージの認証を保証し、盗聴を防止することで、機密性を保持する
2. Authenticate the calling service — Verify the identity of the software sending requests  2. 呼び出し元サービスの認証 - リクエストを送信するソフトウェアのIDを検証する
3. Authorize the service — Using that authenticated identity, check that the action or communication being performed by the service is allowed  3. サービスを認可する - 認証されたIDを使用して、サービスによって実行され るアクションまたはコミュニケーションが許可されていることを確認する
4. Authenticate the end user — Verify the identity of the entity triggering the software to send the request, often a non-person entity (NPE) (e.g., service account, system account)  4. エンドユーザーを認可する - リクエストを送信するソフトウェアのトリガーとなる事業体(多くの場 合、非人間事業体(NPE)(サービスアカウント、システムアカウントなど)のIDを検証する
5. Authorize the end user to resource access — Using the authenticated end-user identity, check that they are allowed to perform the requested action on the target resource  5. エンドユーザーをリソースアクセスに認可する - 認証されたエンドユーザーアイデンティティを使用して、ターゲッ トリソース上で要求されたアクションを実行することが許可されていることを確認する
Achieving a zero-trust runtime requires applying these five controls to all API communications. This guidance further describes additional controls that are necessary for safe and secure API operations beyond identity-based segmentation. These controls should be enforced on all APIs in a system, including those exposed to the outside world (i.e., public APIs) and those intended only for other applications in a given infrastructure (i.e., internal APIs).  ゼロトラストランタイムを達成するには、これら5つの制御をすべてのAPIコミュニケーションに適用する必要がある。本ガイダンスではさらに、ID ベースのセグメンテーションを超えた、安全でセキュアな API 操作に必要な追加のコントロールについて説明する。これらの管理は、外部に公開されるAPI(すなわち公開API)や、特定のインフラストラクチャー内の他のアプリケーショ ンだけを対象とするAPI(すなわち内部API)を含め、システム内のすべてのAPIに対して実施されるべきである。
1.3. API Life Cycle  1.3. APIのライフサイクル
Like all software, APIs grow and change over time as requirements drift and usage patterns change. They also go through a continuous, iterative life cycle, including:  すべてのソフトウェアと同様に、APIは、要求が変化し、使用パターンが変わるにつれて、時間とともに成長し、変化する。APIもまた、以下のような、継続的で反復的なライフサイクルを経る: 
• Plan, Develop, Build, Test, Release — These “pre-runtime” life cycle phases lead to a service that can be deployed in production.  - 計画、開発、ビルド、テスト、リリース - これらの 「プレランタイム 」ライフサイクルのフェーズは、本番環境に展開できるサービスにつながる。
• Deploy, Operate, Monitor, Feedback — These “runtime” life cycle phases involve running and operating a service in production.  - デプロイ、運用、監視、フィードバック - これらの「ランタイム」ライフサイクルフェーズでは、本番環境でのサービスの実行と運用を行う。
DoD Enterprise DevSecOps Fundamentals [1] provides a detailed description of each phase of the software development life cycle. Application of the DevSecOps paradigm in the context of cloud-native applications can be found in [4][5].  DoD Enterprise DevSecOps Fundamentals [1]は、ソフトウェア開発ライフサイクルの各フェーズの詳細な説明を提供する。クラウドネイティブアプリケーションのコンテキストにおける DevSecOps パラダイムの適用については、[4][5]を参照されたい。
20250326-53950
Fig. 3. DevSecOps life cycle phases  図3. DevSecOps ライフサイクルのフェーズ
1.4. Document Goals 1.4. 文書の目標
The goal of this document is to recommend guidance or controls for API protection. These controls are classified into two categories:  この文書の目標は、API 保護のためのガイダンスやコントロールを推奨することである。これらのコントロールは2つのカテゴリーに分類される: 
1. Pre-runtime API protections — These controls need to be applied when designing and building APIs.  1. ランタイム前のAPI防御 - これらの防御は、APIを設計・構築する際に適用する必要がある。
2. Runtime API protections — These controls need to be applied to every API request that an infrastructure serves, not just at the perimeter.  2. ランタイムAPI防御 - これらの防御は、インフラが提供する全てのAPIリクエストに適用される必要がある。
Each of these two categories is further divided into two subcategories based on organizational maturity (i.e., basic and advanced), which enables enterprises to adopt them based on an incremental risk-based approach. これら2つのカテゴリーはそれぞれ、組織の成熟度(ベーシックとアドバンス)に基づいてさらに2つのサブカテゴリーに分けられ、エンタープライズがリスクベースのアプローチに基づいて段階的に導入できるようになっている。 
A prerequisite for defining any API protection measure or policy irrespective of its category or sub-category is that the protections must be expressed in terms of nouns and verbs that pertain to API components, API endpoint components, API requests, and API responses that in turn contain references to resources/data and operations on those resources. These nouns and verbs form the fundamental surface that is exposed to the consumers of APIs and API endpoints. カテゴリーやサブカテゴリーに関係なく、API 保護対策やポリシーを定義するための前提条件は、保護が API コンポーネント、API エンドポイントコンポーネント、API リクエスト、API レスポンスに関連する名詞と動詞で表現されなければならないことである。これらの名詞と動詞は、APIとAPIエンドポイントの消費者に公開される基本的な表面を形成する。 
1.5. Relationship to Other NIST Documents 1.5. 他のNIST文書との関係
Today, most enterprise software development and integration are based on APIs. Section 1.3 articulated the close relationship between software and APIs, demonstrated that API development and deployment follow the same iterative life cycle as the software, and provided NIST guidance on DevSecOps. 今日、ほとんどのエンタープライズソフトウェアの開発と統合はAPIに基づいている。セクション1.3は、ソフトウェアとAPIの密接な関係を明確にし、APIの開発と展開がソフトウェアと同じ反復的なライフサイクルに従うことを示し、DevSecOpsに関するNISTのガイダンスを提供した。 
Another distinguishing feature of the controls recommended for protecting APIs is the capacity to provide assurance for conforming to the principles of zero trust. This is because there is no distinction between internal and external API requests/calls due to the absence of an identifiable network perimeter and the distributed nature of applications on-premises and multiple clouds. This security assurance can be achieved using authentication and authorization controls using identity-based segmentation [2]. Documents that provide recommendations on the configuration of authentication and authorization controls in the context of cloud-native applications (e.g., [2][3]) are also relevant in the context of configuring controls for API protection.  APIを保護するために推奨される防御のもう一つの特徴は、ゼロトラストの原則に適合する保証を提供する能力である。これは、識別可能なネットワーク境界が存在せず、オンプレミスや複数のクラウドに分散したアプリケーションの性質のため、内部と外部のAPIリクエスト/コールの区別がないためである。このセキュリティ保証は、ID ベースのセグメンテーションを使用した認証と認可のコントロールを使用して達成することができる [2]。クラウドネイティブアプリケーションの文脈における認証と認可の制御の構成に関する推奨を提供する文書(例えば、 [2][3])は、API 保護のための制御の構成の文脈にも関連する。
1.6. Document Structure  1.6. 文書構成
This document is organized as follows:  本文書の構成は以下の通りである: 
• Section 2 looks at the risk factors and vulnerabilities associated with APIs and the attack vectors that could exploit those vulnerabilities.  ・セクション 2 は、API に関連するリスク要因と脆弱性、及び、それらの脆弱性を悪用する攻撃ベクターについて見ている。
• Section 3 recommends controls to protect APIs and classifies them into basic and advanced categories that need to be applied prior to runtime or enforced during runtime.  ・セクション 3 は、API を保護するための防御策を推奨し、実行前に適用する必要がある、あるいは実行中に強制 する必要がある、基本的なカテゴリーと高度なカテゴリーに分類する。
• Section 4 provides a detailed analysis of implementation options or patterns for the controls described in Sec. 3 and outlines the advantages and disadvantages of each pattern.  ・セクション4は、セクション3で説明したコントロールの実装オプションやパターンを詳細に分析し,各パターンの長所と短所を概説する。
• Section 5 provides the summary and conclusions.  ・セクション5は、要約と結論を提供する。
• Appendix A provides the classification taxonomy for APIs.  ・附属書Aは、APIの分類分類法を提供する。
• Appendix B illustrates the API controls related to each DevSecOps phase  ・附属書Bは、DevSecOps の各フェーズに関連する API の管理方法を示している。

 

 

| | Comments (0)

2025.02.03

英国 AI安全報告書 2025

こんにちは、丸山満彦です

英国政府が、AIアクションサミットの一環として?、AI安全報告書を公表していますね...

298ページ...

 

GOV. U.K.

・2025.01.29 International AI Safety Report 2025

 

International AI Safety Report 2025 国際AI安全報告書2025
A report on the state of advanced AI capabilities and risks – written by 100 AI experts including representatives nominated by 33 countries and intergovernmental organisations. 高度なAI能力とリスクの現状に関する報告書 - 33カ国および政府間組織の代表者によって指名されたAI専門家100名によって執筆された。
Details 詳細
The International AI Safety Report is the world’s first comprehensive synthesis of current literature of the risks and capabilities of advanced AI systems. Chaired by Turing-award winning computer scientist, Yoshua Bengio, it is the culmination of work by 100 AI experts to advance a shared international understanding of the risks of advanced Artificial Intelligence (AI). 国際AI安全報告書は、高度なAIシステムのリスクと能力に関する現在の文献を総合的にまとめた世界初の包括的な報告書である。チューリング賞受賞者のコンピューター科学者であるヨシュア・ベンジオが議長を務め、100人のAI専門家が高度な人工知能(AI)のリスクに関する国際的な共通理解を深めるために取り組んできた成果の集大成である。
The Chair is supported by an international Expert Advisory Panel made up of representatives from 30 countries, the United Nations (UN), European Union (EU), and Organization for Economic Cooperation and Development (OECD). この議長職は、30カ国、国連(UN)、欧州連合(EU)、経済協力開発機構(OECD)の代表者で構成される国際専門家諮問委員会によって支援されている。
The report does not make policy recommendations. Instead it summarises the scientific evidence on the safety of general-purpose AI to help create a shared international understanding of risks from advanced AI and how they can be mitigated. General-purpose AI – or AI that can perform a wide variety of tasks – is a type of AI that has advanced rapidly in recent years and is widely used by technology companies for a range of consumer and business purposes. この報告書は政策提言を行うものではない。その代わり、汎用AIの安全性に関する科学的証拠をまとめ、高度なAIがもたらすリスクと、その緩和策について国際的な共通理解を促すことを目的としている。汎用AI、すなわち幅広いタスクを実行できるAIは、近年急速に進歩したAIの一種であり、テクノロジー企業が消費者向けおよび企業向けの幅広い用途で広く利用している。
The report is concerned with AI risks and AI safety and focuses on identifying these risks and evaluating methods for mitigating them. It summarises the scientific evidence on 3 core questions – What can general-purpose AI do? What are risks associated with general-purpose AI? and What mitigation techniques are there against these risks? and aims to: 本報告書は、AIのリスクと安全性に焦点を当て、これらのリスクの識別と緩和手法の評価に重点を置いている。本報告書は、3つの主要な疑問点に関する科学的証拠を要約している。すなわち、汎用AIは何ができるのか?汎用AIに関連するリスクとは何か?これらのリスクに対する緩和手法にはどのようなものがあるのか?という3つの疑問点である。本報告書の目的は、
・provide scientific information that will support informed policymaking – it does not recommend specific policies ・情報に基づく政策決定を支援する科学的情報を提供することである。特定の政策を推奨するものではない
・facilitate constructive and evidence-based discussion about the uncertainty of general-purpose AI and its outcomes ・汎用AIの不確実性とその結果について、建設的かつ根拠に基づく議論を促進する
・contribute to an internationally shared scientific understanding of advanced AI safety ・高度なAIの安全性に関する国際的に共有された科学的理解に貢献する
The report was written by a diverse group of academics, guided by world-leading experts in AI. There was no industry or government influence over the content. The secretariat organised a thorough review, which included valuable input from global civil society and industry leaders. The Chair and writers considered all feedback and included it where needed. この報告書は、AI分野の世界的な専門家が指導する多様な学術研究者のグループによって作成された。内容については、業界や政府からの影響は一切ない。事務局は徹底的なレビューを行い、世界中の市民社会や業界のリーダーからの貴重な意見も取り入れた。議長と執筆者はすべてのフィードバックを検討し、必要に応じて取り入れた。
The report will be presented at the AI Action Summit in Paris in February 2025. An interim version of this report was published in May 2024 and presented at the AI Seoul Summit. この報告書は、2025年2月にパリで開催されるAIアクションサミットで発表される。この報告書の暫定版は2024年5月に発表され、AIソウルサミットで提示された。
Background 背景
At the Bletchley AI Safety Summit, 30 nations agreed to build a shared scientific and evidence-based understanding of frontier AI risks through the development of an international, independent, and inclusive report on the risks and capabilities of frontier AI. The International AI Safety Report delivers on that mandate, providing a comprehensive synthesis of the existing literature on the risks and capabilities of advanced AI models. ブレッチリーAI安全サミットでは、30カ国が、最先端AIのリスクと能力に関する国際的、独立した、包括的な報告書の作成を通じて、最先端AIのリスクに関する科学的かつ証拠に基づく共通理解を構築することに合意した。国際AI安全報告書は、その指令に従い、高度なAIモデルのリスクと能力に関する既存の文献を総合的にまとめた包括的な報告書である。
As host of the AI Safety Summit, the UK commissioned Yoshua Bengio to deliver this report, with operational support from a secretariat hosted by the UK AI Safety Institute. The UK government has agreed with participating countries that it will continue to provide the secretariat for the report until a suitable long-term international home is agreed, and Professor Yoshua Bengio will continue acting as Chair for 2025. AI安全サミットの主催国である英国は、ヨシュア・ベンジオに本報告書の作成を依頼し、英国AI安全研究所が主催する事務局が運営面でサポートした。英国政府は、適切な長期的な国際的拠点が合意されるまで、参加国とともに本報告書の事務局を継続して提供することに同意しており、ヨシュア・ベンジオ教授は2025年まで議長を務める予定である。
The report was developed in-line with the principles and procedures, agreed by the Chair and Expert Advisory Panel. 本報告書は、議長および専門家諮問委員会が合意した原則および手順に沿って作成された。

 

・[PDF] International AI Safety Report 2025

20250202-163710

 

・[DOCX] [PDF] 仮訳 

 

目次...

Forewords  まえがき 
About this report  本報告書について 
Update on latest AI advances after the writing of this report: Chair’s note  本報告書執筆後のAIの最新動向:議長による注釈 
Key findings of the report  本報告書の主な調査結果 
Executive Summary  エグゼクティブサマリー 
Introduction  序文 
1. Capabilities of general-purpose AI  1. 汎用AIの能力 
1.1. How general-purpose AI is developed  1.1. 汎用AIの開発方法 
1.2. Current capabilities  1.2. 現在の能力 
1.3. Capabilities in coming years  1.3. 今後数年の能力 
2. Risks  2. リスク 
2.1. Risks from malicious use  2.1. 悪用によるリスク 
2.1.1. Harm to individuals through fake content  2.1.1. 偽のコンテンツによる個人への被害 
2.1.2. Manipulation of public opinion  2.1.2. 世論操作 
2.1.3. Cyber offence  2.1.3. サイバー犯罪 
2.1.4. Biological and chemical attacks  2.1.4. 生物・化学兵器 
2.2. Risks from malfunctions  2.2. 誤作動によるリスク 
2.2.1. Reliability issues  2.2.1. 信頼性の問題 
2.2.2. Bias  2.2.2. バイアス 
2.2.3. Loss of control  2.2.3. 制御不能 
2.3. Systemic risks  2.3. システムリスク 
2.3.1. Labour market risks  2.3.1. 労働市場のリスク 
2.3.2. Global AI R&D divide  2.3.2. グローバルなAI研究開発格差 
2.3.3. Market concentration and single points of failure  2.3.3. 市場の集中と単一障害点 
2.3.4. Risks to the environment  2.3.4. 環境へのリスク 
2.3.5. Risks to privacy  2.3.5. プライバシーへのリスク 
2.3.6. Risks of copyright infringement  2.3.6. 著作権侵害のリスク 
2.4. Impact of open-weight general-purpose AI models on AI risks  2.4. オープンウェイト汎用AIモデルがAIリスクに与える影響 
3. Technical approaches to risk management  3. リスクマネジメントへの技術的アプローチ 
3.1. Risk management overview  3.1. リスクマネジメントの概要 
3.2. General challenges for risk management and policymaking  3.2. リスクマネジメントと政策立案における一般的な課題 
3.2.1. Technical challenges for risk management and policymaking  3.2.1. リスクマネジメントと政策立案における技術的課題 
3.2.2. Societal challenges for risk management and policymaking  3.2.2. リスクマネジメントと政策立案における社会的な課題 
3.3. Risk identification and assessment  3.3. リスクの特定とアセスメント 
3.4. Risk mitigation and monitoring  3.4. リスクの緩和とモニタリング 
3.4.1. Training more trustworthy models  3.4.1. より信頼性の高いモデルのトレーニング 
3.4.2. Monitoring and intervention  3.4.2. モニタリングと介入 
3.4.3. Technical methods for privacy  3.4.3. プライバシーに関する技術的手法 
Conclusion  結論 
List of acronyms  略語一覧 
Glossary  用語集 
How to cite this report  この報告書の引用方法 
References  参考文献 

 

 

主な調査結果

Key findings of the report  主な調査結果 
● The capabilities of general-purpose AI, the type of AI that this report focuses on, have increased rapidly in recent years and have improved further in recent months.[1] A few years ago, the best large language models (LLMs) could rarely produce a coherent paragraph of text. Today, general-purpose AI can write computer programs, generate custom photorealistic images, and engage in extended open-ended conversations. Since the publication of the Interim Report (May 2024), new models have shown markedly better performance at tests of scientific reasoning and programming.  ● 本報告書が注目するAIの種類である汎用AIの能力は、近年急速に向上しており、ここ数カ月でさらに改善されている。[1]数年前までは、どんなに優れた大規模言語モデル(LLM)でも、まとまった段落の文章を作成することはほとんどできなかった。今日、生成的AIはコンピュータ・プログラムを作成し、カスタムメイドの写実的画像を生成し、長時間のオープンエンドな会話をすることができる。中間報告書の発表(2024年5月)以来、新しいモデルは科学的推論とプログラミングのテストにおいて著しく優れた性能を示している。 
● Many companies are now investing in the development of general-purpose AI agents, as a potential direction for further advancement. AI agents are general-purpose AI systems which can autonomously act, plan, and delegate to achieve goals with little to no human oversight. Sophisticated AI agents would be able to, for example, use computers to complete longer projects than current systems, unlocking both additional benefits and additional risks.  ● 多くの企業は現在、さらなる進歩の可能性のある方向として、汎用AIエージェントの開発に投資している。AIエージェントは、自律的に行動し、計画を立て、人間の監視をほとんど受けずに目標を達成するために委任することができる汎用AIシステムである。洗練されたAIエージェントは、例えば、現在のシステムよりも長いプロジェクトを完了するためにコンピュータを使用することができるようになり、新たな利点と新たなリスクの両方を解き放つだろう。 
● Further capability advancements in the coming months and years could be anything from slow to extremely rapid.† Progress will depend on whether companies will be able to rapidly deploy even more data and computational power to train new models, and whether ‘scaling’ models in this way will overcome their current limitations. Recent research suggests that rapidly scaling up models may remain physically feasible for at least several years. But major capability advances may also require other factors: for example, new research breakthroughs, which are hard to predict, or the success of a novel scaling approach that companies have recently adopted.  ● 今後数カ月から数年の間に、さらなる能力の進歩は、ゆっくりとしたものから極めて急速なものまで、あらゆる可能性がある† 進歩は、企業が新しいモデルを訓練するために、さらに多くのデータと計算能力を迅速に展開できるかどうか、また、このようにしてモデルを「スケーリング」することで、現在の限界を克服できるかどうかにかかっている。最近の研究によると、モデルを急速にスケールアップすることは、少なくとも数年間は物理的に可能である。しかし、大きな能力の進歩には、他の要因も必要かもしれない。例えば、予測するのが難しい新たな研究のブレークスルーや、企業が最近採用した斬新なスケーリング・アプローチの成功などである。 
● Several harms from general-purpose AI are already well established. These include scams, non-consensual intimate imagery (NCII) and child sexual abuse material (CSAM), model outputs that are biased against certain groups of people or certain opinions, reliability issues, and privacy violations. Researchers have developed mitigation techniques for these problems, but so far no combination of techniques can fully resolve them. Since the publication of the Interim Report, new evidence of discrimination related to general-purpose AI systems has revealed more subtle forms of bias.  ● 汎用AIによるいくつかの弊害は、すでに十分に確立されている。詐欺、非同意的親密画像(NCII)や児童性的虐待資料(CSAM)、特定のグループや特定の意見にバイアスされたモデル出力、信頼性の問題、プライバシー侵害などである。研究者たちはこれらの問題を緩和する技術を開発してきたが、今のところ、どの技術を組み合わせても完全に解決することはできない。中間報告書の発表以降、汎用AIシステムに関連する識別的な新たな証拠が、より微妙な形のバイアスを明らかにしている。 
● As general-purpose AI becomes more capable, evidence of additional risks is gradually emerging. These include risks such as large-scale labour market impacts, AI-enabled hacking or biological attacks, and society losing control over general-purpose AI. Experts interpret the existing evidence on these risks differently: some think that such risks are decades away, while others think that general-purpose AI could lead to societal-scale harm within the next few years. Recent advances in general-purpose AI capabilities – particularly in tests of scientific reasoning and programming – have generated new evidence for potential risks such as AI-enabled hacking and biological attacks, leading one major AI company to increase its assessment of biological risk from its best model from ‘low’ to ‘medium’.  ● 汎用AIの能力が高まるにつれて、新たなリスクを示す証拠が徐々に明らかになってきている。これには、大規模な労働市場への影響、AIを利用したハッキングや生物学的攻撃、社会が汎用AIを制御できなくなるといったリスクが含まれる。専門家の間では、こうしたリスクに関する既存の証拠の解釈が異なっており、そうしたリスクは数十年先のことだと考える人もいれば、汎用AIが今後数年以内に社会規模の被害をもたらす可能性があると考える人もいる。汎用AIの能力、特に科学的推論やプログラミングのテストにおける最近の進歩は、AIを利用したハッキングや生物学的攻撃といった潜在的リスクに関する新たな証拠を生み出し、ある大手AI企業は、最良のモデルによる生物学的リスクの評価を「低」から「中」に引き上げるに至った。 
● Risk management techniques are nascent, but progress is possible. There are various technical methods to assess and reduce risks from general-purpose AI that developers can employ and regulators can require, but they all have limitations. For example, current interpretability techniques for explaining why a general-purpose AI model produced any given output remain severely limited. However, researchers are making some progress in addressing these limitations. In addition, researchers and policymakers are increasingly trying to standardise risk management approaches, and to coordinate internationally.  ● リスクマネジメント技術は発展途上だが、進歩は可能だ。汎用AIがもたらすリスクをアセスメントし、低減するための技術的手法には、開発者が採用し、規制当局が求めることができる様々なものがあるが、いずれも限界がある。例えば、汎用AIモデルがどのような出力を生成したかを説明するための現在の解釈可能性技術は、依然として大きく制限されている。しかし、研究者たちはこれらの限界に対処するために、ある程度の進歩を遂げつつある。さらに、研究者や政策立案者は、リスクマネジメントのアプローチを標準化し、国際的な協調を図ろうとしている。 
● The pace and unpredictability of advancements in general-purpose AI pose an ‘evidence dilemma’ for policymakers.[2] Given sometimes rapid and unexpected advancements, policymakers will often have to weigh potential benefits and risks of imminent AI advancements without having a large body of scientific evidence available. In doing so, they face a dilemma. On the one hand, pre-emptive risk mitigation measures based on limited evidence might turn out to be ineffective or unnecessary. On the other hand, waiting for stronger evidence of impending risk could leave society unprepared or even make mitigation impossible – for instance if sudden leaps in AI capabilities, and their associated risks, occur. Companies and governments are developing early warning systems and risk management frameworks that may reduce this dilemma. Some of these trigger specific mitigation measures when there is new evidence of risks, while others require developers to provide evidence of safety before releasing a new model.  ● 汎用AIの進歩のペースと予測不可能性は、政策立案者に「証拠のジレンマ」をもたらす。[2]時として急速で予期せぬ進歩を遂げるため、政策立案者は、科学的根拠が十分でないまま、目前に迫ったAIの進歩がもたらす潜在的な利益とリスクを秤にかけなければならないことが多い。その際、政策立案者はジレンマに直面する。一方では、限られた証拠に基づく先制的なリスク緩和策は、効果がない、あるいは不要であることが判明するかもしれない。他方で、差し迫ったリスクについてより強力な証拠を待つことは、社会に備えを怠らせたり、緩和を不可能にする可能性さえある。企業やガバナンス政府は、このジレンマを軽減する可能性のある早期警告システムやリスクマネジメントの枠組みを開発している。これらの中には、 、リスクに関する新たな証拠が発見された場合に特定の緩和措置を発動するものもあれば、新モデルをリリースする前に安全性の証拠を提供するよう開発者に求めるものもある。 
● There is broad consensus among researchers that advances regarding the following questions would be helpful: How rapidly will general-purpose AI capabilities advance in the coming years, and how can researchers reliably measure that progress? What are sensible risk thresholds to trigger mitigations? How can policymakers best gain access to information about general-purpose AI that is relevant to public safety? How can researchers, technology companies, and governments reliably assess the risks of general-purpose AI development and deployment? How do general-purpose AI models work internally? How can general-purpose AI be designed to behave reliably?  ● 研究者の間では、以下のような疑問に関する進歩が有用であるという点で幅広いコンセンサスが得られている:今後数年間で、汎用AIの能力はどの程度急速に進歩するのか、また研究者はどのようにしてその進歩を確実に測定できるのか。緩和のきっかけとなるリスクのしきい値はどの程度か?政策立案者は、公共の安全に関連する汎用AIに関する情報に、どのようにアクセスすればよいのか?研究者、テクノロジー企業、政府は、汎用AIの開発・展開のリスクをどのように確実にアセスメントできるのか?汎用AIのモデルは内部でどのように機能するのか?汎用AIが確実に動作するように設計するにはどうすればよいか? 
● AI does not happen to us: choices made by people determine its future. The future of general-purpose AI technology is uncertain, with a wide range of trajectories appearing to be possible even in the near future, including both very positive and very negative outcomes. This uncertainty can evoke fatalism and make AI appear as something that happens to us. But it will be the decisions of societies and governments on how to navigate this uncertainty that determine which path we will take. This report aims to facilitate constructive and evidence-based discussion about these decisions. ● AIは私たちに起こるものではない。人々の選択がその未来を決めるのだ。汎用AI技術の未来は不確実であり、非常にポジティブな結果も非常にネガティブな結果も含め、近い将来でさえ幅広い軌道が可能であるように見える。この不確実性は宿命論を呼び起こし、AIを私たちに起こることのように見せる可能性がある。しかし、この不確実性をどのように乗り切るかについての社会や政府の決断こそが、われわれがどのような道を歩むかを決めるのである。本報告書は、こうした意思決定について、建設的で証拠に基づいた議論を促進することを目的としている。

 

エグゼクティブサマリー...

Executive Summary  エグゼクティブサマリー 
The purpose of this report  本報告書の目的 
This report synthesises the state of scientific understanding of general-purpose AI – AI that can perform a wide variety of tasks – with a focus on understanding and managing its risks.  本報告書は、汎用AI(さまざまなタスクを実行できるAI)の科学的理解の現状を、そのリスクの理解とマネジメントに焦点を当てながらまとめたものである。 
This report summarises the scientific evidence on the safety of general-purpose AI. The purpose of this report is to help create a shared international understanding of risks from advanced AI and how they can be mitigated. To achieve this, this report focuses on general-purpose AI – or AI that can perform a wide variety of tasks – since this type of AI has advanced particularly rapidly in recent years and has been deployed widely by technology companies for a range of consumer and business purposes. The report synthesises the state of scientific understanding of general-purpose AI, with a focus on understanding and managing its risks.  本報告書は、汎用AIの安全性に関する科学的根拠をまとめたものである。本報告書の目的は、高度なAIがもたらすリスクとその緩和方法について、国際的な共通理解を得ることにある。そのために本報告書では、汎用AI、つまり多種多様なタスクを実行できるAIに焦点を当てている。なぜなら、この種のAIは近年特に急速に進歩しており、テクノロジー企業によって消費者やビジネスのさまざまな目的で広く展開されているからである。本報告書は、汎用AIの科学的理解の現状を、そのリスクの理解とマネジメントに焦点を当てて総括している。 
Amid rapid advancements, research on general-purpose AI is currently in a time of scientific discovery, and – in many cases – is not yet settled science. The report provides a snapshot of the current scientific understanding of general-purpose AI and its risks. This includes identifying areas of scientific consensus and areas where there are different views or gaps in the current scientific understanding.  急速な進歩の中で、汎用AIの研究は現在、科学的発見の時期にあり、多くの場合、まだ科学として確立していない。本報告書は、汎用AIとそのリスクに関する現在の科学的理解のスナップショットを提供する。これには、科学的コンセンサスのある分野と、現在の科学的理解に異なる見解やギャップがある分野の特定も含まれる。 
People around the world will only be able to fully enjoy the potential benefits of general-purpose AI safely if its risks are appropriately managed. This report focuses on identifying those risks and evaluating technical methods for assessing and mitigating them, including ways that general-purpose AI itself can be used to mitigate risks. It does not aim to comprehensively assess all possible societal impacts of general-purpose AI. Most notably, the current and potential future benefits of general-purpose AI – although they are vast – are beyond this report’s scope. Holistic policymaking requires considering both the potential benefits of general-purpose AI and the risks covered in this report. It also requires taking into account that other types of AI have different risk/benefit profiles compared to current general-purpose AI.  汎用AIのリスクが適切にマネジメントされて初めて、世界中の人々が汎用AIの潜在的な恩恵を安全に十分に享受できるようになる。本報告書は、汎用AIそのものがリスクを緩和するために利用できる方法を含め、そうしたリスクを特定し、それをアセスメントし緩和するための技術的手法を評価することに焦点を当てている。汎用AIの社会的影響の可能性をすべて包括的に評価することを目的としているわけではない。最も注目すべきは、汎用AIの現在および将来の潜在的な利益は、膨大なものではあるが、本報告書の範囲を超えていることである。総合的な政策立案には、汎用AIの潜在的なメリットと、本報告書で取り上げたリスクの両方を考慮する必要がある。また、他の種類のAIは、現在の汎用AIとは異なるリスク/ベネフィット・プロファイルを持つことも考慮する必要がある。 
The three main sections of the report summarise the scientific evidence on three core questions: What can general-purpose AI do? What are risks associated with general-purpose AI? And what mitigation techniques are there against these risks?  報告書の3つの主要セクションは、3つの核となる疑問に関する科学的証拠を要約している:汎用AIは何ができるのか? 汎用AIにはどのようなリスクがあるのか?そして、これらのリスクに対してどのような緩和技術があるのか? 
Section 1 – Capabilities of general-purpose AI: What can general-purpose AI do now and in the future?  第1節 汎用AIの能力:汎用AIは現在、そして将来何ができるのか? 
General purpose AI capabilities have improved rapidly in recent years, and further advancements could be anything from slow to extremely rapid. 
一般的な-目的別AIの能力は近年急速に向上しており、今後の改善はゆっくりとしたものから極めて急速なものまである。 
What AI can do is a key contributor to many of the risks it poses, and according to many metrics, general-purpose AI capabilities have been progressing rapidly. Five years ago, the leading general-purpose AI language models could rarely produce a coherent paragraph of text. Today, some general-purpose AI models can engage in conversations on a wide range of topics, write computer programs, or generate realistic short videos from a description. However, it is technically challenging to reliably estimate and describe the capabilities of general-purpose AI.  AIに何ができるかは、AIがもたらすリスクの多くに重要な寄与をしており、多くの指標によれば、汎用AIの能力は急速に進歩している。5年前、主要な汎用AI言語モデルは、まとまった段落の文章をほとんど作成できなかった。今日、いくつかの汎用AIモデルは、幅広いトピックについて会話したり、コンピューター・プログラムを書いたり、説明からリアルな短い動画を生成したりすることができる。しかし、汎用AIの能力を確実に推定し、説明することは技術的に困難である。 
AI developers have rapidly improved the capabilities of general-purpose AI in recent years, mostly through ‘scaling’.[1] They have continually increased the resources used for training new models (this is often referred to as ‘scaling’) and refined existing approaches to use those resources more efficiently. For example, according to recent estimates, state-of-the-art AI models have seen annual increases of approximately 4x in computational resources ('compute') used for training and 2.5x in training dataset size.  AI開発者は近年、汎用AIの能力を急速に改善してきたが、そのほとんどは「スケーリング」によるものだ。[1]新しいモデルの学習に使用するリソースを継続的に増やし(これはしばしば「スケーリング」と呼ばれる)、それらのリソースをより効率的に使用するために既存のアプローチを改良してきた。例えば、最近の推定によると、最先端のAIモデルは、学習に使用される計算リソース(「コンピュート」)が年間約4倍、学習データセットサイズが2.5倍増加している。 
The pace of future progress in general-purpose AI capabilities has substantial implications for managing emerging risks, but experts disagree on what to expect even in the coming months and years. Experts variously support the possibility of general-purpose AI capabilities advancing slowly, rapidly, or extremely rapidly.  汎用AI能力の将来的な進歩のペースは、新たなリスクのマネジメントに大きな意味を持つが、今後数カ月から数年の間にも何が予想されるかについては、専門家の間でも意見が分かれている。専門家は、汎用AIの能力がゆっくりと、急速に、あるいは極めて急速に進歩する可能性を様々に支持している。 
Experts disagree about the pace of future progress because of different views on the promise of further ‘scaling’ – and companies are exploring an additional, new type of scaling that might further accelerate capabilities. While scaling has often overcome the limitations of previous systems, experts disagree about its potential to resolve the remaining limitations of today’s systems, such as unreliability at acting in the physical world and at executing extended tasks on computers. In recent months, a new type of scaling has shown potential for further improving capabilities: rather than just scaling up the resources used for training models, AI companies are also increasingly interested in ‘inference scaling’ – letting an already trained model use more computation to solve a given problem, for example to improve on its own solution, or to write so-called ‘chains of thought’ that break down the problem into simpler steps.  専門家の間では、さらなる「スケーリング」の有望性に対する見解の相違から、将来の進歩のペースについて意見が分かれており、企業は能力をさらに加速させる可能性のある、追加の新しいタイプのスケーリングを模索している(†)。スケーリングはしばしば以前のシステムの限界を克服してきたが、専門家の間では、物理的な世界で行動する際の信頼性の低さや、コンピュータ上で長時間のタスクを実行する際の信頼性の低さなど、今日のシステムに残っている限界を解決する可能性については意見が分かれている。ここ数カ月、新しいタイプのスケーリングが能力をさらに向上させる可能性を示している。AI企業は、モデルの学習に使用するリソースを単にスケールアップするだけでなく、「推論スケーリング」にもますます関心を寄せている。これは、すでに学習されたモデルに、与えられた問題を解くためにより多くの計算を使用させるもので、例えば、自身の解法を改善したり、問題をより単純なステップに分解するいわゆる「思考の連鎖」を記述したりすることができる。 
Several leading companies that develop general-purpose AI are betting on ‘scaling’ to continue leading to performance improvements. If recent trends continue, by the end of 2026 some general-purpose AI models will be trained using roughly 100x more training compute than 2023's most compute-intensive models, growing to 10,000x more training compute by 2030, combined with algorithms that achieve greater capabilities for a given amount of available computation. In addition to this potential scaling of training resources, recent trends such as inference scaling and using models to generate training data could mean that even more compute will be used overall. However, there are potential bottlenecks to further increasing both data and compute rapidly, such as the availability of data, AI chips, capital, and local energy capacity. Companies developing general-purpose AI are working to navigate these potential bottlenecks.  汎用AIを開発するいくつかの大手企業は、「スケーリング」が性能改善につながり続けることに賭けている。最近のトレンドが続けば、2026年末までに、汎用AIモデルのいくつかは、2023年の最も計算負荷の高いモデルよりもおよそ100倍多い学習用計算量を使用して学習され、2030年までに10,000倍多い学習用計算量を使用するようになり、利用可能な計算量が与えられた場合に、より高い能力を達成するアルゴリズムと組み合わされる。このようなトレーニングリソースのスケーリングの可能性に加え、推論のスケーリングやトレーニングデータを生成するためにモデルを使用するなどの最近のトレンドは、全体としてさらに多くの計算機が使用されることを意味する可能性がある。しかし、データ、AIチップ、資本、地域のエネルギー容量の利用可能性など、データと計算の両方をさらに急速に増加させるには、潜在的なボトルネックがある。汎用AIを開発している企業は、これらの潜在的なボトルネックを回避するために取り組んでいる。 
Since the publication of the Interim Report (May 2024), general-purpose AI has reached expert-level performance in some tests and competitions for scientific reasoning and programming, and companies have been making large efforts to develop autonomous AI agents. Advances in science and programming have been driven by inference scaling techniques such as writing long ‘chains of thought’. New studies suggest that further scaling such approaches, for instance allowing models to analyse problems by writing even longer chains of thought than today’s models, could lead to further advances in domains where reasoning matters more, such as science, software engineering, and planning. In addition to this trend, companies are making large efforts to develop more advanced general-purpose AI agents, which can plan and act autonomously to work towards a given goal. Finally, the market price of using general-purpose AI of a given capability level has dropped sharply, making this technology more broadly accessible and widely used.  中間報告発表(2024年5月)以降、汎用AIは科学的推論やプログラミングに関するいくつかのテストやコンテストでエキスパートレベルの性能に達し、企業は自律型AIエージェントの開発に大きく力を入れている。科学とプログラミングの進歩は、長い「思考の連鎖」を書くなどの推論スケーリング技術によって推進されてきた。新たな研究は、このようなアプローチのスケーリングをさらに進め、例えば、 、現在のモデルよりもさらに長い思考の連鎖を書くことによって問題を分析できるようにすれば、科学、ソフトウェア工学、プランニングなど、推論がより重要な領域でさらなる進歩が期待できることを示唆している。このトレンドに加え、企業は、より高度な汎用AIエージェントの開発に大きな努力を払っている。このエージェントは、与えられた目標に向かって自律的に計画を立て、行動することができる。最終的に、ある能力レベルの汎用AIを使用する市場価格は急激に低下し、この技術がより広く利用できるようになり、広く利用されるようになった。 
This report focuses primarily on technical aspects of AI progress, but how fast general-purpose AI will advance is not a purely technical question. The pace of future advancements will also depend on non-technical factors, potentially including the approaches that governments take to regulating AI. This report does not discuss how different approaches to regulation might affect the speed of development and adoption of general-purpose AI.  本報告書は主にAIの進歩の技術的側面に焦点を当てているが、汎用AIの進歩の速度は純粋に技術的な問題ではない。将来の進歩のペースは、非技術的な要因にも左右され、政府がAIを規制する際のアプローチも含まれる可能性がある。本報告書では、規制に対するアプローチの違いが、汎用AIの開発・普及のスピードにどのような影響を与えるかについては論じていない。 
Section 2 – Risks: What are risks associated with general-purpose AI?  第2節 リスク:汎用AIに伴うリスクとは何か? 
Several harms from general-purpose AI are already well-established. As general-purpose AI becomes more capable, evidence of additional risks is gradually emerging.  汎用AIによるいくつかの弊害はすでに確立されている。汎用AIの能力が高まるにつれて、新たなリスクの証拠が徐々に明らかになってきている。 
This report classifies general-purpose AI risks into three categories: malicious use risks; risks from malfunctions; and systemic risks. Each of these categories contains risks that have already materialised as well as risks that might materialise in the next few years.  本報告書では、汎用AIのリスクを「悪意のある利用リスク」「誤作動によるリスク」「システムリスク」の3カテゴリーに分類している。これらのカテゴリーにはそれぞれ、すでに顕在化しているリスクと、今後数年間に顕在化する可能性のあるリスクが含まれている。 
Risks from malicious use: malicious actors can use general-purpose AI to cause harm to individuals, organisations, or society. Forms of malicious use include:  意のある利用によるリスク:悪意のある行為者は、汎用AIを利用して個人、組織、社会に危害を加える可能性がある。悪意のある利用の形態には以下のようなものがある: 
Harm to individuals through fake content: Malicious actors can currently use general-purpose AI to generate fake content that harms individuals in a targeted way. These malicious uses include non-consensual 'deepfake' pornography and AI-generated CSAM, financial fraud through voice impersonation, blackmail for extortion, sabotage of personal and professional reputations, and psychological abuse. However, while incident reports of harm from AI-generated fake content are common, reliable statistics on the frequency of these incidents are still lacking.  偽コンテンツによる個人への被害:悪意ある行為者は現在、生成的AIを利用して、個人を標的に害する偽コンテンツを生成することができる。このような悪意のある利用には、非合意の「ディープフェイク」ポルノやAIが生成したCSAM、音声なりすましによる金銭詐欺、恐喝のための恐喝、個人的・職業的評判の妨害、心理的虐待などが含まれる。しかし、AIが生成的な偽コンテンツによる被害のインシデント報告はよくあるものの、こうしたインシデントの頻度に関する信頼できる統計はまだ不足している。 
● Manipulation of public opinion: General-purpose AI makes it easier to generate persuasive content at scale. This can help actors who seek to manipulate public opinion, for instance to affect political outcomes. However, evidence on how prevalent and how effective such efforts are remains limited. Technical countermeasures like content watermarking, although useful, can usually be circumvented by moderately sophisticated actors.  ● 世論を操作する:生成的AIによって、説得力のあるコンテンツを大規模に生成することが容易になる。これは、例えば政治的な結果に影響を与えるために世論を操作しようとする行為者を助けることができる。しかし、そのような取り組みがどれほど普及し、どれほど効果的であるかについての証拠はまだ限られている。コンテンツの電子透かしのような技術的な対策は、有用ではあるが、中程度に洗練された行為者であれば通常回避することができる。 
● Cyber offence: General-purpose AI can make it easier or faster for malicious actors of varying skill levels to conduct cyberattacks. Current systems have demonstrated capabilities in low- and medium-complexity cybersecurity tasks, and state-sponsored actors are actively exploring AI to survey target systems. New research has confirmed that the capabilities of general-purpose AI related to cyber offence are significantly advancing, but it remains unclear whether this will affect the balance between attackers and defenders.  ● サイバー攻撃汎用AIは、さまざまなスキルレベルの悪意ある行為者がサイバー攻撃を行うことを容易に、あるいは迅速にすることができる。現在のシステムは、低・中複雑度のサイバーセキュリティ・タスクで能力を発揮しており、国家的支援を受けた行為者は、標的システムを調査するためのAIを積極的に模索している。新たな研究では、サイバー攻撃に関する汎用AIの能力が著しく進歩していることが確認されているが、これが攻撃側と防御側のバランスに影響を与えるかどうかはまだ不明である。 
● Biological and chemical attacks: Recent general-purpose AI systems have displayed some ability to provide instructions and troubleshooting guidance for reproducing known biological and chemical weapons and to facilitate the design of novel toxic compounds. In new experiments that tested for the ability to generate plans for producing biological weapons, a general-purpose AI system sometimes performed better than human experts with access to the internet. In response, one AI company increased its assessment of biological risk from its best model from ‘low’ to ‘medium’. However, real-world attempts to develop such weapons still require substantial additional resources and expertise. A comprehensive assessment of biological and chemical risk is difficult because much of the relevant research is classified.  ● 生物・化学兵器による攻撃最近の汎用AIシステムは、既知の生物・化学兵器を再現するための指示やトラブルシューティングのガイダンスを提供したり、新規毒性化合物の設計を容易にしたりする能力をある程度示している。生物兵器を製造するための計画を生成する能力をテストした新しい実験では、生成的AIシステムが、インターネットにアクセスできる人間の専門家よりも優れた結果を出すこともあった。これを受けて、あるAI企業は、最良のモデルによる生物学的リスクのアセスメントを「低」から「中」に引き上げた。しかし、このような兵器を現実の世界で開発しようとする試みには、まだかなりの追加資源と専門知識が必要である。生物学的・化学的リスクの包括的アセスメントは、関連研究の多くが機密扱いであるため困難である。 
Since the publication of the Interim Report, general-purpose AI has become more capable in domains that are relevant for malicious use. For example, researchers have recently built general-purpose AI systems that were able to find and exploit some cybersecurity vulnerabilities on their own and, with human assistance, discover a previously unknown vulnerability in widely used software. General-purpose AI capabilities related to reasoning and to integrating different types of data, which can aid research on pathogens or in other dual-use fields, have also improved.  中間報告書の発表以来、汎用AIは悪意のある利用に関連する領域でより高い能力を発揮するようになっている。例えば、研究者たちは最近、サイバーセキュリティの脆弱性を自ら発見して悪用することができる汎用AIシステムを構築し、また、人間の支援を得て、広く使われているソフトウェアのこれまで知られていなかった脆弱性を発見した。推論や、病原体やその他の二次利用分野の研究を助けることができる、さまざまな種類のデータの統合に関連する汎用AIの能力も改善されている。 
Risks from malfunctions: general-purpose AI can also cause unintended harm. Even when users have no intention to cause harm, serious risks can arise due to the malfunctioning of general-purpose AI. Such malfunctions include:  誤作動によるリスク:汎用AIは意図しない危害をもたらすこともある。ユーザーに危害を加える意図がない場合でも、汎用AIの誤作動によって深刻なリスクが生じる可能性がある。そのような誤作動には以下のようなものがある: 
● Reliability issues: Current general-purpose AI can be unreliable, which can lead to harm. For example, if users consult a general-purpose AI system for medical or legal advice, the system might generate an answer that contains falsehoods. Users are often not aware of the limitations of an AI product, for example due to limited ‘AI literacy’, misleading advertising, or miscommunication. There are a number of known cases of harm from reliability issues, but still limited evidence on exactly how widespread different forms of this problem are.  ● 信頼性の問題:現在の汎用AIは信頼性が低く、それが弊害につながる可能性がある。例えば、ユーザーが医療や法律に関する助言を求めて汎用AIシステムに相談した場合、システムが虚偽を含む回答を生成する可能性がある。例えば、「AIリテラシー」の低さ、誤解を招くような広告、誤ったコミュニケーションなどが原因で、ユーザーはAI製品の限界を認識していないことが多い。信頼性の問題による危害の事例は数多く知られているが、この問題のさまざまな形態がどの程度広まっているかについては、正確な証拠はまだ限られている。 
● Bias: General-purpose AI systems can amplify social and political biases, causing concrete harm. They frequently display biases with respect to race, gender, culture, age, disability, political opinion, or other aspects of human identity. This can lead to discriminatory outcomes including unequal resource allocation, reinforcement of stereotypes, and systematic neglect of underrepresented groups or viewpoints. Technical approaches for mitigating bias and discrimination in general-purpose AI systems are advancing, but face trade-offs between bias mitigation and competing objectives such as accuracy and privacy, as well as other challenges.  ● バイアス:汎用AIシステムは、社会的・政治的バイアスを増幅し、具体的な害をもたらす可能性がある。AIシステムは、人種、性別、文化、年齢、障害、政治的意見、あるいは人間のアイデンティティの他の側面に関してバイアスを示すことが多い。これは、不平等な資源配分、固定観念の強化、代表的でないグループや視点の組織的無視など、差別的な結果につながる可能性がある。汎用AIシステムにおけるバイアスと差別を緩和する技術的アプローチは進歩しているが、バイアスの緩和と、精度やプライバシーなど競合する目的との間のトレードオフや、その他の課題に直面している。 
● Loss of control: ‘Loss of control’ scenarios are hypothetical future scenarios in which one or more general-purpose AI systems come to operate outside of anyone's control, with no clear path to regaining control. There is broad consensus that current general-purpose AI lacks the capabilities to pose this risk. However, expert opinion on the likelihood of loss of control within the next several years varies greatly: some consider it implausible, some consider it likely to occur, and some see it as a modest-likelihood risk that warrants attention due to its high potential severity. Ongoing empirical and mathematical research is gradually advancing these debates.  ● 制御の喪失:「制御の喪失」シナリオとは、1つまたは複数の全般統制AIシステムが、誰の制御も及ばず、制御を取り戻す明確な道筋もないまま動作するようになる仮想的な未来シナリオである。現在の汎用AIには、このようなリスクを引き起こす能力がないことは広くコンセンサスが得られている。しかし、今後数年以内に制御不能になる可能性に関する専門家の意見は大きく異なる。ある者はあり得ないと考え、ある者は発生する可能性が高いと考え、またある者は潜在的な重大性が高いため注意を要する中程度の可能性のリスクと見ている。現在進行中の実証的・数学的研究が、こうした議論を徐々に進展させている。 
Since the publication of the Interim Report, new research has led to some new insights about risks of bias and loss of control. The evidence of bias in general-purpose AI systems has increased, and recent work has detected additional forms of AI bias. Researchers have observed modest further advancements towards AI capabilities that are likely necessary for commonly discussed loss of control scenarios to occur. These include capabilities for autonomously using computers, programming, gaining unauthorised access to digital systems, and identifying ways to evade human oversight.  中間報告書の発表以来、新たな研究により、バイアスと制御不能のリスクに関する新たな洞察がいくつか得られている。汎用AIシステムにおけるバイアスの証拠が増加し、最近の研究では、新たな形態のAIバイアスが検知された。研究者たちは、一般的に議論されている制御不能のシナリオが発生するために必要と思われるAIの能力について、さらに緩やかな進歩を観察している。これには、コンピュータを自律的に使用する能力、プログラミング能力、デジタルシステムへの不正アクセス能力、人間の監視を逃れる方法の特定能力などが含まれる。 
Systemic risks: beyond the risks directly posed by capabilities of individual models, widespread deployment of general-purpose AI is associated with several broader systemic risks. Examples of systemic risks range from potential labour market impacts to privacy risks and environmental effects:  システミックリスク:個々のモデルの能力が直接もたらすリスクを超えて、汎用AIの広範な展開は、いくつかのより広範なシステミックリスクと関連している。システミックリスクの例は、潜在的な労働市場への影響からプライバシーリスクや環境への影響まで多岐にわたる: 
● Labour market risks: General-purpose AI, especially if it continues to advance rapidly, has the potential to automate a very wide range of tasks, which could have a significant effect on the labour market. This means that many people could lose their current jobs. However, many economists expect that potential job losses could be offset, partly or potentially even completely, by the creation of new jobs and by increased demand in non-automated sectors.  ● 労働市場のリスク:汎用AIは、特にそれが急速に進歩し続ければ、非常に幅広い作業を自動化する可能性を秘めており、労働市場に大きな影響を与える可能性がある。つまり、多くの人々が現在の職を失う可能性がある。しかし、多くのエコノミストは、潜在的な雇用損失は、新たな雇用の創出や非自動化分野での需要増加によって、部分的に、あるいは潜在的に完全に相殺される可能性があると予想している。 
● Global AI R&D divide: General-purpose AI research and development (R&D) is currently concentrated in a few Western countries and China. This ‘AI divide’ has the potential to increase much of the world’s dependence on this small set of countries. Some experts also expect it to contribute to global inequality. The divide has many causes, including a number of causes that are not unique to AI. However, in significant part it stems from differing levels of access to the very expensive compute needed to develop general-purpose AI: most low- and middle-income countries (LMICs) have significantly less access to compute than high-income countries (HICs).  グローバルなAI研究開発格差:汎用AIの研究開発(R&D)は現在、少数の欧米諸国と中国に集中している。この「AI格差」は、世界の多くの国々がこの少数の国々への依存度を高める可能性がある。一部の専門家は、これが世界的な不平等に寄与する可能性もあると予測している。この格差には多くの原因があり、AIに特有ではない原因もいくつかある。しかし、その大きな要因は、汎用AIの開発に必要な非常に高価なコンピューティングへのアクセスレベルの違いから生じている。ほとんどの低・中所得国(LMIC)は、高所得国(HIC)よりもコンピューティングへのアクセスが大幅に少ない。
● Market concentration and single points of failure: A small number of companies currently dominate the market for general-purpose AI. This market concentration could make societies more vulnerable to several systemic risks. For instance, if organisations across critical sectors, such as finance or healthcare, all rely on a small number of general-purpose AI systems, then a bug or vulnerability in such a system could cause simultaneous failures and disruptions on a broad scale.  ● 市場の集中と単一障害点:現在、少数の企業が汎用AIの市場を独占している。このような市場の集中は、社会をいくつかのシステミックリスクに対してより脆弱にする可能性がある。例えば、金融や医療など重要なセクターの組織がすべて少数の汎用AIシステムに依存している場合、そのようなシステムにバグや脆弱性があれば、大規模な障害や混乱を同時に引き起こす可能性がある。 
● Environmental risks: Growing compute use in general-purpose AI development and deployment has rapidly increased the amounts of energy, water, and raw material consumed in building and operating the necessary compute infrastructure. This trend shows no clear indication of slowing, despite progress in techniques that allow compute to be used more efficiently. General-purpose AI also has a number of applications that can either benefit or harm sustainability efforts.  ● 環境リスク:汎用AIの開発・展開におけるコンピュート利用の拡大により、必要なコンピュート・インフラを構築・運用するために消費されるエネルギー、水、原材料の量が急速に増加している。この傾向は、コンピュートのより効率的な利用を可能にする技術の進歩にもかかわらず、減速する明確な兆候は見られない。汎用AIにも、持続可能性への取り組みに有益にも有害にもなりうるアプリケーションが数多くある。 
● Privacy risks: General-purpose AI can cause or contribute to violations of user privacy. For example, sensitive information that was in the training data can leak unintentionally when a user interacts with the system. In addition, when users share sensitive information with the system, this information can also leak. But general-purpose AI can also facilitate deliberate violations of privacy, for example if malicious actors use AI to infer sensitive information about specific individuals from large amounts of data. However, so far, researchers have not found evidence of widespread privacy violations associated with general-purpose AI.  ● プライバシーリスク:汎用AIは、ユーザーのプライバシー侵害を引き起こしたり、助長したりする可能性がある。例えば、学習データに含まれていた機密情報が、ユーザーがシステムと対話する際に意図せず漏れる可能性がある。さらに、ユーザーが機密情報をシステムと共有した場合、その情報も漏れる可能性がある。しかし、汎用AIは意図的なプライバシー侵害を助長する可能性もある。例えば、悪意のある行為者がAIを使って大量のデータから特定の個人に関する機密情報を推測する場合などだ。しかし、これまでのところ、研究者は汎用AIに関連する広範なプライバシー侵害の証拠を発見していない。 
● Copyright infringements: General-purpose AI both learns from and creates works of creative expression, challenging traditional systems of data consent, compensation, and control. Data collection and content generation can implicate a variety of data rights laws, which vary across jurisdictions and may be under active litigation. Given the legal uncertainty around data collection practices, AI companies are sharing less information about the data they use. This opacity makes third-party AI safety research harder.  ● 著作権侵害:汎用AIは創造的表現の作品から学習し、また創造的表現の作品を創作し、データ同意、補償、管理に関する従来の制度に挑戦する。データ収集とコンテンツ生成は、様々なデータ権利法に関わる可能性があり、それらは司法管轄区域によって異なり、訴訟も活発に行われている可能性がある。データ収集慣行をめぐる法的不確実性を考慮すると、AI企業は自社が使用するデータに関する情報をあまり共有しないようになっている。この不透明さが、サードパーティによるAIの安全性研究を難しくしている。 
Since the publication of the Interim Report, additional evidence on the labour market impacts of general-purpose AI has emerged, while new developments have heightened privacy and copyrights concerns. New analyses of labour market data suggest that individuals are adopting general-purpose AI very rapidly relative to previous technologies. The pace of adoption by businesses varies widely by sector. In addition, recent advances in capabilities have led to general-purpose AI being deployed increasingly in sensitive contexts such as healthcare or workplace monitoring, which creates new privacy risks. Finally, as copyright disputes intensify and technical mitigations to copyright infringements remain unreliable, data rights holders have been rapidly restricting access to their data.  中間報告書の発表以降、汎用AIの労働市場への影響に関する新たな証拠が登場する一方で、新たな進展によりプライバシーや著作権に関する懸念が高まっている。労働市場データの新たな分析によれば、個人が汎用AIを採用するスピードは、従来のテクノロジーと比較して非常に速いことが示唆されている。企業による採用のペースは分野によって大きく異なる。加えて、最近の機能の進歩により、ヘルスケアや職場のモニタリングなど、センシティブな場面での汎用AIの展開が増えており、新たなプライバシーリスクが発生している。最後に、著作権紛争が激化し、著作権侵害に対する技術的緩和が依然として信頼できないため、データ権利者はデータへのアクセスを急速に制限している。 
Open-weight models: an important factor in evaluating many risks that a general-purpose AI model might pose is how it is released to the public. So-called ‘open-weight models’ are AI models whose central components, called ‘weights’, are shared publicly for download. Open-weight access facilitates research and innovation, including in AI safety, as well as increasing transparency and making it easier for the research community to detect flaws in models. However, open-weight models can also pose risks, for example by facilitating malicious or misguided use that is difficult or impossible for the developer of the model to monitor or mitigate. Once model weights are available for public download, there is no way to implement a wholesale rollback of all existing copies or ensure that all existing copies receive safety updates. Since the Interim Report, high-level consensus has emerged that risks posed by greater AI openness should be evaluated in terms of ‘marginal’ risk: the extent to which releasing an open-weight model would increase or decrease a given risk, relative to risks posed by existing alternatives such as closed models or other technologies.  オープンウエイトモデル:汎用AIモデルがもたらすであろう多くのリスクを評価する上で重要な要素は、それがどのように一般に公開されているかである。いわゆる「オープンウエイトモデル」とは、「ウエイト」と呼ばれる中心的なコンポーネントがダウンロードできるように公開されているAIモデルのことである。オープンウェイトへのアクセスは、AIの安全性を含む研究とイノベーションを促進し、透明性を高め、研究コミュニティがモデルの欠陥を検出しやすくする。しかし、オープン・ウェイト・モデルは、例えば、モデルの開発者が監視したり緩和したりすることが困難または不可能な悪意ある利用や誤った利用を助長するなどのリスクをもたらす可能性もある。一旦モデルの重みが一般にダウンロードできるようになると、既存の全てのコピーを全面的にロールバックしたり、既存の全てのコピーが安全性の更新を受けることを保証したりする方法はない。中間報告書以降、AIのオープン性を高めることによってもたらされるリスクは、「限界」リスク、すなわち、クローズドモデルや他の技術といった既存の代替案がもたらすリスクと比較して、オープンウエイトモデルを公開することによって所定のリスクがどの程度増減するかという観点から評価されるべきであるというハイレベルなコンセンサスが生まれている。 
Section 3 – Risk management: What techniques are there for managing risks from general-purpose AI?  第3節 リスクマネジメント:汎用AIのリスクマネジメントにはどのような手法があるのか? 
Several technical approaches can help manage risks, but in many cases the best available approaches still have highly significant limitations and no quantitative risk estimation or guarantees that are available in other safety-critical domains.  いくつかの技術的アプローチはリスクマネジメントに役立つが、多くの場合、利用可能な最善のアプローチにはまだ非常に大きな限界があり、他のセーフティ・クリティカルな領域で利用可能な定量的なリスク評価や保証はない。 
Risk management – identifying and assessing risks, and then mitigating and monitoring them – is difficult in the context of general-purpose AI. Although risk management has also been highly challenging in many other domains, there are some features of general-purpose AI that appear to create distinctive difficulties.  リスクの特定とアセスメント、そして緩和とモニタリングといったリスクマネジメントは、汎用AIの文脈では難しい。リスクマネジメントは他の多くの領域でも非常に難しいが、汎用AIには独特の難しさを生み出す特徴がある。 
Several technical features of general-purpose AI make risk management in this domain particularly difficult. They include, among others:  汎用AIのいくつかの技術的特徴が、この分野のリスクマネジメントを特に難しくしている。それらには特に以下のようなものがある: 
● The range of possible uses and use contexts for general-purpose AI systems is unusually broad. For example, the same system may be used to provide medical advice, analyse computer code for vulnerabilities, and generate photos. This increases the difficulty of comprehensively anticipating relevant use cases, identifying risks, or testing how systems will behave in relevant real-world circumstances.  ● 汎用AIシステムの可能な用途と使用コンテクストの範囲は、異常に広い。例えば、同じシステムを医療アドバイスに使ったり、コンピュータコードの脆弱性を分析したり、写真を生成したりすることもできる。このため、関連するユースケースを包括的に予測したり、リスクを特定したり、関連する実世界の状況でシステムがどのように振る舞うかをテストしたりすることが難しくなっている。 
● Developers still understand little about how their general-purpose AI models operate. This lack of understanding makes it more difficult both to predict behavioural issues and to explain and resolve known issues once they are observed. Understanding remains elusive mainly because general-purpose AI models are not programmed in the traditional sense.  ● 開発者たちは、汎用のAIモデルがどのように動作するかについて、まだほとんど理解していない。この理解不足は、行動上の問題を予測することも、既知の問題が観察された後にそれを説明し解決することも難しくしている。主に、汎用AIモデルは伝統的な意味でプログラムされていないため、理解はつかみにくいままだ。 
Instead, they are trained: AI developers set up a training process that involves a large volume of data, and the outcome of that training process is the general-purpose AI model. The inner workings of these models are largely inscrutable, including to the model developers. Model explanation and ‘interpretability’ techniques can improve researchers’ and developers’ understanding of how general-purpose AI models operate, but, despite recent progress, this research remains nascent.  その代わり、AIは訓練される:AI開発者は、大量のデータを含むトレーニングプロセスを設定し、そのトレーニングプロセスの結果が汎用AIモデルである。これらのモデルの内部構造は、モデル開発者を含め、ほとんど不可解である。モデルの説明と「解釈可能性」技術は、汎用AIモデルがどのように動作するかについての研究者や開発者の理解を改善することができるが、最近の進歩にもかかわらず、この研究はまだ始まったばかりである。 
● Increasingly capable AI agents – general-purpose AI systems that can autonomously act, plan, and delegate to achieve goals – will likely present new, significant challenges for risk management. AI agents typically work towards goals autonomously by using general software such as web browsers and programming tools. Currently, most are not yet reliable enough for widespread use, but companies are making large efforts to build more capable and reliable AI agents and have made progress in recent months. AI agents will likely become increasingly useful, but may also exacerbate a number of the risks discussed in this report and introduce additional difficulties for risk management. Examples of such potential new challenges include the possibility that users might not always know what their own AI agents are doing, the potential for AI agents to operate outside of anyone’s control, the potential for attackers to ‘hijack’ agents, and the potential for AI-to-AI interactions to create complex new risks. Approaches for managing risks associated with agents are only beginning to be developed.  ● AIエージェント(目標達成のために自律的に行動し、計画し、委任することができる汎用AIシステム)の能力が向上すれば、リスクマネジメントに新たな重大な課題をもたらす可能性が高い。AIエージェントは通常、ウェブ・ブラウザやプログラミング・ツールなどの一般的なソフトウェアを使用することで、自律的に目標に向かって行動する。現在のところ、そのほとんどはまだ広く使用できるほど信頼できるものではないが、企業はより有能で信頼性の高いAIエージェントを構築するために大きな努力を払っており、ここ数カ月で進歩を遂げている。AIエージェントは今後ますます便利になっていくだろうが、本報告書で取り上げた多くのリスクを悪化させ、リスクマネジメントに新たな困難をもたらす可能性もある。そのような潜在的な新たな課題の例としては、ユーザーが自身のAIエージェントが何をしているのか常に把握しているとは限らない可能性、AIエージェントが誰のコントロールも及ばないところで動作する可能性、攻撃者がエージェントを「ハイジャック」する可能性、AI同士の相互作用が複雑な新たなリスクを生み出す可能性などが挙げられる。エージェントに関連するリスクをマネジメントするアプローチは、まだ開発され始めたばかりである。 
Besides technical factors, several economic, political, and other societal factors make risk management in the field of general-purpose AI particularly difficult.  技術的要因に加え、経済的、政治的、その他の社会的要因も、汎用AIの分野におけるリスクマネジメントをとりわけ難しくしている。 
● The pace of advancement in general-purpose AI creates an 'evidence dilemma' for decision-makers.[2] Rapid capability advancement makes it possible for some risks to emerge in leaps; for example, the risk of academic cheating using general-purpose AI shifted from negligible to widespread within a year. The more quickly a risk emerges, the more difficult it is to manage the risk reactively and the more valuable preparation becomes. However, so long as evidence for a risk remains incomplete, decision-makers also cannot know for sure whether the risk will emerge or perhaps even has already emerged. This creates a trade-off:  ● 汎用AIの進歩の速さは、意思決定者に「証拠のジレンマ」をもたらす。[2]例えば、汎用AIを使った学歴詐称のリスクは、1年以内に無視できるものから広まるものへと変化した。リスクが急速に顕在化すればするほど、そのリスクを反応的にマネジメントすることは難しくなり、準備の価値は高まる。しかし、リスクの証拠が不完全なままである限り、意思決定者はリスクが顕在化するかどうか、あるいはおそらくすでに顕在化しているかどうかさえも確実に知ることはできない。これはトレードオフの関係にある: 
implementing pre-emptive or early mitigation measures might prove unnecessary, but waiting for conclusive evidence could leave society vulnerable to risks that emerge rapidly. Companies and governments are developing early warning systems and risk management frameworks that may reduce this dilemma. Some of these trigger specific mitigation measures when there is new evidence of risks, while others require developers to provide evidence of safety before releasing a new model.  しかし、決定的な証拠を待つことは、急速に顕在化するリスクに対する脆弱性を社会に残すことになりかねない。企業や政府( )は、このジレンマを軽減する可能性のある早期警告システムやリスクマネジメント枠組みを開発している。これらの中には、リスクに関する新たな証拠が発見された場合に特定の緩和措置を発動するものもあれば、新モデルをリリースする前に安全性の証拠を提供するよう開発者に求めるものもある。 
● There is an information gap between what AI companies know about their AI systems and what governments and non-industry researchers know. Companies often share only limited information about their general-purpose AI systems, especially in the period before they are widely released. Companies cite a mixture of commercial concerns and safety concerns as reasons to limit information sharing. However, this information gap also makes it more challenging for other actors to participate effectively in risk management, especially for emerging risks.  ● AI企業が自社のAIシステムについて知っていることと、政府や非産業研究者が知っていることの間には、情報格差がある。企業は汎用AIシステムについて、特に広くリリースされる前の時期には、限られた情報しか共有しないことが多い。企業は、情報共有を制限する理由として、商業的な懸念と安全上の懸念が混在していることを挙げている。しかし、このような情報格差は、他のアクターがリスクマネジメントに効果的に参加すること、特に新たなリスクについて参加することを難しくしている。 
● Both AI companies and governments often face strong competitive pressure, which may lead them to deprioritise risk management. In some circumstances, competitive pressure may incentivise companies to invest less time or other resources into risk management than they otherwise would. Similarly, governments may invest less in policies to support risk management in cases where they perceive trade-offs between international competition and risk reduction.  ● AI企業も政府も、しばしば強い競争圧力に直面し、リスクマネジメントを軽視するようになるかもしれない。状況によっては、競争圧力は、企業がリスクマネジメントに投資する時間やその他のリソースを、そうでない場合よりも少なくする動機付けになるかもしれない。同様に、政府は、国際競争とリスク削減との間にトレードオフがあると認識する場合には、リスクマネジメントを支援する政策への投資を少なくする可能性がある。 
Nonetheless, there are various techniques and frameworks for managing risks from general-purpose AI that companies can employ and regulators can require. These include methods for identifying and assessing risks, as well as methods for mitigating and monitoring them.  とはいえ、汎用AIによるリスクマネジメントには、企業が採用し、規制当局が求めることのできる様々な手法や枠組みがある。これには、リスクを特定・アセスメントする手法や、リスクを緩和・監視する手法が含まれる。 
● Assessing general-purpose AI systems for risks is an integral part of risk management, but existing risk assessments are severely limited. Existing evaluations of general-purpose AI risk mainly rely on ‘spot checks’, i.e. testing the behaviour of a general-purpose AI in a set of specific situations. This can help surface potential hazards before deploying a model. However, existing tests often miss hazards and overestimate or underestimate general-purpose AI capabilities and risks, because test conditions differ from the real world.  ● 汎用AIシステムのリスクアセスメントはリスクマネジメントの不可欠な要素であるが、既存のリスクアセスメントは極めて限定的である。既存の汎用AIのリスク評価は、主に「抜き打ち検査」、つまり特定の状況における汎用AIの挙動テストに頼っている。これは、モデルを展開する前に潜在的な危険を表面化させるのに役立つ。しかし、既存のテストでは、テスト条件が現実世界と異なるため、ハザードを見逃したり、汎用AIの能力やリスクを過大評価したり過小評価したりすることが多い。 
● For risk identification and assessment to be effective, evaluators need substantial expertise, resources, and sufficient access to relevant information. Rigorous risk assessment in the context of general-purpose AI requires combining multiple evaluation approaches. These range from technical analyses of the models and systems themselves to evaluations of possible risks from certain use patterns. Evaluators need substantial expertise to conduct such evaluations correctly. For comprehensive risk assessments, they often also need more time, more direct access to the models and their training data, and more information about the technical methodologies used than the companies developing general-purpose AI typically provide.  ● リスクの特定とアセスメントを効果的に行うためには、評価者は相当な専門知識、リソース、関連情報への十分なアクセスを必要とする。汎用AIの文脈における厳密なリスクアセスメントには、複数の評価アプローチを組み合わせる必要がある。これらは、モデルやシステム自体の技術的分析から、特定の使用パターンから起こりうるリスクの評価まで多岐にわたる。評価者は、このような評価を正しく行うために相当な専門知識を必要とする。包括的なリスクアセスメントを行うためには、汎用AIの開発企業が通常提供するよりも多くの時間、モデルやその学習データへの直接アクセス、使用されている技術的方法論に関する多くの情報も必要となることが多い。 
● There has been progress in training general-purpose AI models to function more safely, but no current method can reliably prevent even overtly unsafe outputs. For example, a technique called ‘adversarial training’ involves deliberately exposing AI models to examples designed to make them fail or misbehave during training, aiming to build resistance to such cases. However, adversaries can still find new ways ('attacks') to circumvent these safeguards with low to moderate effort. In addition, recent evidence suggests that current training methods – which rely heavily on imperfect human feedback – may inadvertently incentivise models to mislead humans on difficult questions by making errors harder to spot. Improving the quantity and quality of this feedback is an avenue for progress, though nascent training techniques using AI to detect misleading behaviour also show promise.  ● 汎用のAIモデルをより安全に機能させるためのトレーニングは進歩しているが、現在の手法では、あからさまに安全でない出力さえも確実に防ぐことはできない。例えば、「敵対的トレーニング」と呼ばれる手法では、トレーニング中にAIモデルをわざと失敗させたり誤動作させたりするようなサンプルにさらすことで、そのようなケースに対する耐性を構築することを目的としている。しかし、敵対者は、これらのセーフガードを回避する新しい方法(「攻撃」)を、低~中程度の労力で見つけることができる。加えて、不完全な人間のフィードバックに大きく依存している現在のトレーニング方法は、エラーを発見しにくくすることで、難しい問題で人間を惑わすようなインセンティブをモデルに不用意に与えている可能性があることを、最近の証拠が示唆している。このようなフィードバックの量と質を改善することは、進歩の道であるが、誤解を招く行動を検知するためにAIを使用する新しいトレーニング技術も有望である。 
● Monitoring – identifying risks and evaluating performance once a model is already in use – and various interventions to prevent harmful actions can improve the safety of a general-purpose AI after it is deployed to users. Current tools can detect AI-generated content, track system performance, and identify potentially harmful inputs/outputs, though moderately skilled users can often circumvent these safeguards. Several layers of defence that combine technical monitoring and intervention capabilities with human oversight improve safety but can introduce costs and delays. In the future, hardware-enabled mechanisms could help customers and regulators to monitor general-purpose AI systems more effectively during deployment and potentially help verify agreements across borders, but reliable mechanisms of this kind do not yet exist.  ● モニタリング(モデルがすでに使用された後のリスクの特定と性能評価)、および有害な行動を防止するための様々な介入は、汎用AIがユーザーに展開された後の安全性を向上させることができる。生成的なツールは、AIが生成したコンテンツを検知し、システムのパフォーマンスを追跡し、潜在的に有害な入出力を特定することができるが、熟練したユーザーであれば、これらのセーフガードを回避できることも多い。技術的な監視・介入機能と人間による監視を組み合わせた何層もの防御は、安全性を向上させるが、コストと遅延をもたらす可能性がある。将来的には、ハードウェアに対応したメカニズムが、顧客や規制当局が展開中の汎用AIシステムをより効果的に監視するのに役立ち、国境を越えた合意を検証するのに役立つ可能性があるが、この種の信頼できるメカニズムはまだ存在しない。 
● Multiple methods exist across the AI lifecycle to safeguard privacy. These include removing sensitive information from training data, model training approaches that control how much information is learned from data (such as ‘differential privacy’ approaches), and techniques for using AI with sensitive data that make it hard to recover the data (such as ‘confidential computing’ and other privacy-enhancing technologies). Many privacy-enhancing methods from other research fields are not yet applicable to general-purpose AI systems due to the computational requirements of AI systems. In recent months, privacy protection methods have expanded to address AI's growing use in sensitive domains including smartphone assistants, AI agents, always-listening voice assistants, and use in healthcare or legal practice.  ● AIのライフサイクルには、プライバシーを保護するための複数の方法が存在する。これには、学習データから機密情報を除去する方法、データから学習する情報量を制御するモデル・トレーニングの方法(「差分プライバシー」アプローチなど)、データの復元を困難にする機密データをAIで使用する技術(「機密コンピューティング」やその他のプライバシー強化技術など)などがある。他の研究分野のプライバシー向上手法の多くは、AIシステムの計算要件のため、汎用AIシステムにはまだ適用できない。ここ数カ月で、プライバシー保護手法は、スマートフォンアシスタント、AIエージェント、常時聞き手の音声アシスタント、医療や法律実務での使用など、AIがセンシティブな領域で使用されるようになっていることに対応するために拡大している。 
Since the publication of the Interim Report, researchers have made some further progress towards being able to explain why a general-purpose AI model has produced a given output. Being able to explain AI decisions could help manage risks from malfunctions ranging from bias and factual inaccuracy to loss of control. In addition, there have been growing efforts to standardise assessment and mitigation approaches around the world.  中間報告書の発表以来、研究者たちは、汎用AIモデルがなぜある出力を出したのかを説明できるようになるために、さらなる進歩を遂げてきた。AIの判断を説明できるようになれば、バイアスや事実誤認から制御不能に至るまで、誤作動によるリスクマネジメントに役立つ可能性がある。さらに、世界中でアセスメントと緩和のアプローチを標準化する取り組みが活発化している。 
Conclusion: A wide range of trajectories for the future of general-purpose AI are possible, and much will depend on how societies and governments act  結論:汎用AIの未来にはさまざまな軌跡が考えられるが、その多くは社会や政府がどのように行動するかにかかっている。 
The future of general-purpose AI is uncertain, with a wide range of trajectories appearing possible even in the near future, including both very positive and very negative outcomes. But nothing about the future of general-purpose AI is inevitable. How general-purpose AI gets developed and by whom, which problems it gets designed to solve, whether societies will be able to reap general-purpose AI’s full economic potential, who benefits from it, the types of risks we expose ourselves to, and how much we invest into research to manage risks – these and many other questions depend on the choices that societies and governments make today and in the future to shape the development of general-purpose AI.  汎用AIの未来は不確実であり、非常にポジティブな結果も非常にネガティブな結果も含め、近い将来にも幅広い軌跡があり得ると思われる。しかし、汎用AIの未来に必然性はない。汎用AIが誰によってどのように開発されるのか、どのような問題を解決するために設計されるのか、社会は汎用AIの経済的可能性をフルに享受できるのか、誰がその恩恵を受けるのか、私たちはどのようなリスクにさらされるのか、リスクを管理するための研究にどの程度投資するのか--こうした疑問や他の多くの疑問は、汎用AIの開発を形成するために社会や政府が今日および将来行う選択にかかっている。 
To help facilitate constructive discussion about these decisions, this report provides an overview of the current state of scientific research and discussion on managing the risks of general-purpose AI.  こうした決定に関する建設的な議論を促進するため、本報告書では、汎用AIのリスクマネジメントに関する科学的研究と議論の現状を概観する。 
The stakes are high. We look forward to continuing this effort.  賭け金は高い。私たちはこの努力を続けることを楽しみにしている。 

 

 


 

まるちゃんの情報セキュリティ気まぐれ日記

AI Action Summit

・2024.11.29 米国 AI安全研究所 安全保障等に関わる政府タスクフォースの設立、AI安全機構国際ネットワークの設立総会で初代議長に、NIST AI 100-4の発表 (2024.11.20)

2024.05.23 外務省 岸田総理大臣のAIソウル・サミット首脳セッションへの参加(ビデオメッセージ)+ AIソウル・サミット首脳セッション出席国による安全、革新的で包摂的なAIのためのソウル宣言 

 

 

 

| | Comments (0)

2025.01.15

書籍 NISC編 サイバーセキュリティ関係法令Q&Aハンドブック

こんにちは、丸山満彦です。

商事法務から、「サイバーセキュリティ関係法令Q&Aハンドブック」内閣官房内閣サイバーセキュリティセンター(NISC)編が発刊されましたね。。。

Img_7843

 

これは、NISCの委員会で取りまとめたものを書籍化したものです。

https://security-portal.nisc.go.jp/guidance/law_handbook.html

 

私も作成に関わっていますが、 岡村委員長をはじめ、ワーキンググループの方、事務局がかなりご苦労をされてできたものです。。。

裁判例を含めて関連情報が厚いのが良いところですかね。。。そして、内閣官房謹製。。。

手元に1冊あると参考になることも多いし、PDFと違った良さもあるので、ぜひ1冊。。。

 

●2025.01.10「サイバーセキュリティ関係法令Q&Aハンドブック」内閣官房内閣サイバーセキュリティセンター(NISC)編

商事法務

全国官報販売協同組合

Amazon

e-hon

楽天book

 


まるちゃんの情報セキュリティ気まぐれ日記

・2023.09.26 内閣官房 NISC 関係法令Q&Aハンドブック Ver 2.0

・2020.03.03 NISC 「サイバーセキュリティ関係法令Q&Aハンドブック 」について

 

| | Comments (0)

より以前の記事一覧