« 米国 NIST SP 1334( 初期公開ドラフト) OT 環境における可搬保管媒体のサイバーセキュリティリスクの軽減 (2025.07.15) | Main | 米国 ホワイトハウス AI行動計画(2025.07.23) »

2025.07.25

欧州委員会 汎用 AI 実践規範 (2025.07.10)

こんにちは、丸山満彦です。

欧州委員会から、汎用AI実践規範(第1章「透明性」と第2章「著作権」第3章「安全とセキュリティ」)が公表されていますね...

第1章「透明性」と第2章「著作権」は汎用 AI モデルのすべてのプロバイダを対象としtています。第3章「安全とセキュリティ」は、最先端のモデルを提供する限られたプロバイダのみを対象としていますね...

 

European Commission

・2025.07.10 General-Purpose AI Code of Practice now available

General-Purpose AI Code of Practice now available 汎用 AI 実践規範が公表された
The European Commission has received the final version of the General-Purpose AI Code of Practice, a voluntary tool developed by 13 independent experts, with input from over 1,000 stakeholders, including model providers, small and medium-sized enterprises, academics, AI safety experts, rightsholders, and civil society organisations. 欧州委員会は、13 人の独立専門家が、モデルプロバイダ、中小企業、学者、AI 安全の専門家、権利者、市民団体など 1,000 以上の利害関係者の意見を取り入れて策定した自主的なツールである「汎用 AI 実践規範」の最終版を受け取った。
The Code is designed to help industry comply with the AI Act's rules on general-purpose AI, which will enter into application on 2 August 2025. The rules become enforceable by the AI Office of the Commission one year later as regards new models and two years later as regards existing models. This aims to ensure that general-purpose AI models placed on the European market — including the most powerful ones — are safe and transparent. この規範は、2025年8月2日に施行される AI 法における汎用 AI に関する規則の遵守を支援することを目的としている。この規則は、新規モデルについては 1 年後、既存モデルについては 2 年後に、欧州委員会の AI 事務局によって施行される。これは、欧州市場に上市される汎用 AI モデル(最も強力なモデルを含む)の安全性と透明性を確保することを目的としている。
The Code consists of three chapters: Transparency and Copyright, both addressing all providers of general-purpose AI models, and Safety and Security, relevant only to a limited number of providers of the most advanced models. この規範は 3 章で構成されており、第 1 章「透明性」と第 2 章「著作権」は汎用 AI モデルのすべてのプロバイダを対象とし、第 3 章「安全とセキュリティ」は、最先端のモデルを提供する限られたプロバイダのみを対象としている。
Once the Code is endorsed by the Member States and the Commission, providers of general-purpose AI models who voluntarily sign the Code will be able to demonstrate compliance with the relevant AI Act obligations by adhering to the Code. In doing so, signatories to the Code will benefit from a reduced administrative burden and increased legal certainty compared to providers that prove compliance in other ways. この規範が加盟国および欧州委員会によって承認されると、この規範に自主的に署名した汎用 AI モデルのプロバイダは、この規範を順守することで、関連する AI 法の義務を順守していることを実証することができる。これにより、この規範の署名者は、他の方法で順守を証明するプロバイダに比べ、行政負担の軽減と法的確実性の向上というメリットを享受することができる。
More information on the Code of Practice can be found in the press release online. この規範の詳細については、オンラインのプレスリリースを参照。
Find further information about: 以下の情報も参照すること:
The Code of Practice on GPAI 汎用AIに関する実践規範
Questions and Answers on the Code of Practice on GPAI 汎用AIに関する実践規範に関する質問と回答
More information about the drawing-up process of the GPAI Code of Practice 汎用AI実践規範の策定プロセスに関する詳細情報

 


実践規範

The General-Purpose AI Code of Practice

The General-Purpose AI Code of Practice 汎用AI実践規範
The Code of Practice helps industry comply with the AI Act legal obligations on safety, transparency and copyright of general-purpose AI models. この実践規範は、汎用AIモデルの安全性、透明性、著作権に関するAI法の法的義務に業界が準拠するための支援を目的としている。
The General-Purpose AI (GPAI) Code of Practice is a voluntary tool, prepared by independent experts in a multi-stakeholder process, designed to help industry comply with the AI Act’s obligations for providers of general-purpose AI models. Read more about the timeline and the drafting process of the code. 汎用 AI(GPAI)実践規範は、多国間協議プロセスにおいて独立した専門家によって作成された自主的なツールであり、業界が汎用 AI モデルのプロバイダに対する AI 法の義務を遵守するのを支援することを目的としている。規範のスケジュールおよび作成プロセスについて詳しくは、こちらを参照。
The code was published on July 10, 2025. In the following weeks, Member States and the Commission will assess its adequacy. Additionally, the code will be complemented by Commission guidelines on key concepts related to general-purpose AI models, to be published still in July. この規範は 2025 年 7 月 10 日に公表された。今後数週間のうちに、加盟国および欧州委員会がその妥当性を評価する。さらに、この規範は、汎用 AI モデルに関する重要な概念に関する欧州委員会のガイドラインによって補完され、7 月中に公表される予定である。
After the code is endorsed by Member States and the Commission, AI model providers who voluntarily sign it can show they comply with the AI Act by adhering to the code. This will reduce their administrative burden and give them more legal certainty than if they proved compliance through other methods. 加盟国および欧州委員会によってこの規範が承認された後、自主的にこの規範に署名した AI モデルプロバイダは、この規範を順守することで AI 法を遵守していることを示すことができる。これにより、他の方法で順守を証明する場合に比べ、行政負担が軽減され、法的確実性が高まる。
Find more information on the questions and answers (Q&A) about the code of practice for General-Purpose AI. 汎用 AI 実践規範に関する質問と回答(Q&A)の詳細については、こちらを参照。
Providers of general-purpose AI models may sign the code by completing the Signatory Form and sending the signed form to [mail]. Potential Signatories may also email this address to ask any questions on the process of signing up to the code. Read more information about the signature process in this this dedicated Q&A. 汎用 AI モデルのプロバイダは、署名者フォームに記入し、署名したフォームを [mail]に送信することで、この規範に署名することができる。署名を検討している方は、このアドレスに、実践規範への署名手続きに関する質問を E メールで送付すること。署名手続きの詳細については、この専用 Q&A を参照。
The three chapters of the code 実践規範の 3 つの章
Below you can download the code, consisting of three separately authored chapters: Transparency, Copyright, and Safety and Security. 以下から、3 つの章(透明性、著作権、安全とセキュリティ)で構成される実践規範をダウンロードできる。
The Chapters on Transparency and Copyright offer all providers of general-purpose AI models a way to demonstrate compliance with their obligations under Article 53 AI Act. 透明性および著作権に関する章では、汎用 AI モデルのすべてのプロバイダが、AI 法第 53 条に基づく義務の遵守を実証する方法を紹介している。
The Chapters on Safety and Security is only relevant to the small number of providers of the most advanced models, those that are subject to the AI Act's obligations for providers of general-purpose AI models with systemic risk under Article 55 AI Act. 安全およびセキュリティに関する章は、AI 法第 55 条に基づく、システムリスクのある汎用 AI モデルのプロバイダに対する AI 法の義務の対象となる、ごく少数の最先端モデルのプロバイダにのみ関係する。
Transparency 透明性
The Transparency chapter (PDF) offers a user-friendly Model Documentation Form (DOCX) which allows providers to easily document the information necessary to comply with the AI Act obligation to on model providers to ensure sufficient transparency. 透明性に関する章(PDF)では、プロバイダが AI 法におけるモデルプロバイダに対する十分な透明性の確保義務を遵守するために必要な情報を簡単に文書化できる、ユーザーフレンドリーなモデル文書フォーム(DOCX)を提供している。
Copyright 著作権
The Copyright chapter (PDF) offers providers practical solutions to meet the AI Act's obligation to put in place a policy to comply with EU copyright law. 著作権に関する章(PDF)では、EU の著作権法を遵守するためのポリシーを策定するという AI 法の義務をプロバイダが満たすための実践的なソリューションを紹介している。
Safety and Security 安全とセキュリティ
The Safety and Security chapter (PDF) outlines concrete state-of-the-art practices for managing systemic risks, i.e. risks from the most advanced models. Providers can rely on this chapter to comply with the AI Act obligations for providers of general-purpose AI models with systemic risk. 「安全とセキュリティ」の章(PDF)では、システムリスク、すなわち最先端のモデルから生じるリスクを管理するための、具体的な最先端の慣行の概要を紹介している。プロバイダは、この章を参考にして、システムリスクのある汎用 AI モデルのプロバイダに対する AI 法の義務を遵守することができる。

 

・[PDF] Transparency chapter 

20250723-83421

 

・[PDF] Copyright chapter

20250723-83457

 

 

・[PDF] Safety and Security chapter

20250723-83539

 

3つまとめて...

・[DOCX][PDF] 仮訳

フォーム

・[DOCX] 原文 仮訳

 

 


策定

・2025.07.14 Drawing-up a General-Purpose AI Code of Practice

Drawing-up a General-Purpose AI Code of Practice 汎用 AI 実践規範の策定
The first General-Purpose AI Code of Practice received by the Commission details the AI Act rules for providers of general-purpose AI models and general-purpose AI models with systemic risks. 欧州委員会が最初に受け取った汎用 AI 実践規範には、汎用 AI モデルおよびシステムリスクを伴う汎用 AI モデルのプロバイダに対する AI 法規則の詳細が記載されている。
The European AI Office facilitated the drawing-up of the , chaired by independent experts, involving nearly 1000 stakeholders, as well as EU Member States representatives, European and international observers. 欧州 AI 事務局は、独立した専門家が議長を務め、1,000 近くの利害関係者、EU 加盟国の代表者、欧州および国際的なオブザーバーが参加して、この規範の策定を支援した。
Why a Code of Practice for General-Purpose AI? 汎用 AI 実践規範が必要な理由
General-purpose AI (GPAI) models can perform a wide range of tasks and are becoming the basis for many AI systems in the EU. Some of these models could carry systemic risks if they are very capable or widely used. To ensure safe and transparent AI, the AI Act puts in place rules for providers of such models. This includes transparency and copyright-related rules. For models that may carry systemic risks, providers should assess and mitigate these risks. 汎用 AI(GPAI)モデルは、幅広いタスクを実行でき、EU の多くの AI システムの基盤となりつつある。これらのモデルの中には、非常に高性能であるか、広く使用されている場合、システムリスクをもたらす可能性があるものもある。安全で透明性の高い AI を確保するため、AI 法は、このようなモデルの提供者に規則を定めている。これには、透明性および著作権に関する規則が含まれる。システムリスクをもたらす可能性のあるモデルについては、提供者はこれらのリスクを評価し、緩和すべきである。
The AI Act rules on general-purpose AI apply from 2 August 2025. The AI Office has been facilitating the drawing-up of a Code of Practice to detail out these rules. 汎用 AI に関する AI 法規則は、2025 年 8 月 2 日から適用される。AI 事務局は、これらの規則の詳細を規定した実践規範の策定を推進している。
The Code represents a voluntary tool prepared by independent experts designed to help industry comply with the AI Act’s rules on general-purpose AI.  この規範は、業界が汎用 AI に関する AI 法規則を順守するのを支援するために、独立した専門家によって作成された自主的なツールである。
How has the Code of Practice been drawn-up? 実践規範はどのように策定されたのか?
he Code has been prepared in an iterative drafting process. On 30 September, the AI Office hosted the Kick-off Event for the Code of Practice Plenary involving nearly 1000 participants, including a high number of professional organisations. These are the eligible respondents to the call to participate in the drawing-up launched by the AI Office on 30 July 2024. この規範は、反復的なドラフト作成プロセスを経て作成された。9 月 30 日、AI 事務局は、実践規範に関する本会議のキックオフイベントを開催し、多くの専門機関を含む 1,000 人近くの参加者が集まった。これらは、2024 年 7 月 30 日に AI 事務局が募集した、実践規範の策定への参加の呼びかけに応募した者たちだ。
The plenary was structured in four working groups on specific topics. Following the kick-off Plenary, the participants were convene three times virtually for drafting rounds with discussions organised in working groups. For each of the working groups, discussions were facilitated by chairs and vice-chairs, who were selected and appointed by the AI Office from interested independent experts. Participants were able express comments during each of those meetings or within two weeks in writing. 本会議は、特定のテーマに関する 4 つの作業部会で構成された。キックオフ本会議の後、参加者は 3 回、作業部会での議論によるドラフト作成のためのオンライン会議を開催した。各作業部会では、AI 事務局が関心のある独立した専門家の中から選任した議長と副議長が議論を進行した。参加者は、各会議で、あるいは 2 週間以内に書面で意見を表明することができた。
The Code has been drafted in an inclusive and transparent process, with a variety of interested stakeholders involved: general-purpose AI model providers, downstream providers, industry organisations, civil society, rightsholders and other entities, as well as academia and independent experts. この規範は、汎用 AI モデルプロバイダ、下流プロバイダ、業界団体、市民社会、権利者、その他の事業体、学界、独立専門家など、さまざまな利害関係者が参加し、包括的かつ透明性の高いプロセスを経てドラフトが作成された。
As the main addressees of the Code, providers of general-purpose AI models were invited to dedicated workshops with the Chairs and Vice-Chairs, contributing to informing each iterative drafting round, in addition to their Plenary participation. The AI Office has ensured transparency into these discussions. この規範の主な対象者である汎用 AI モデルプロバイダは、本会議への参加に加え、議長および副議長による特別ワークショップにも招待され、各ドラフト作成の反復作業に情報を提供した。AI 事務局は、これらの議論の透明性を確保した。
Close involvement of EU Member States representatives has been ensured throughout the process, via the AI Board. The AI Office also invited other public bodies and agencies from all over the world working on risk assessment and mitigation for general-purpose AI models to the Plenary as observers. EU 加盟国の代表者は、AI 委員会を通じて、プロセス全体に密接に関与した。AI 事務局は、汎用 AI モデルのリスクアセスメントおよび緩和に取り組む世界中の他の公的機関や機関も、オブザーバーとして本会議に招待した。
The AI Office played a pivotal role throughout the process. By facilitating the drawing-up, coordinating the discussions and documenting the outcomes, the AI Office guaranteed that the Chairs, Vice-Chairs, and all the participants in the Plenary had access to the information and can contribute meaningfully. This oversight helps maintain an open and collaborative environment, fostering trust and accountability in the development of the Code of Practice. AI事務局はプロセス全体で中心的役割を果たした。草案の作成を促進し、議論を調整し、結果を文書化することで、AI事務局は議長、副議長、および本会議のすべての参加者が情報にアクセスし、意味のある貢献ができるよう保証した。この監督は、実践規範の開発において開かれた協働的な環境を維持し、信頼と責任感を育むのに役立つ。
Template for the summary of training data トレーニングデータの要約テンプレート
In parallel to the Code of practice process, the AI Office is also developing a template on the sufficiently detailed summary of training data that general-purpose AI model providers are required to make public according to Article 53(1)d) of the AI Act. The AI Office collected input from stakeholders during a broader multi-stakeholder consultation on general-purpose AI, where more than 430 responses were provided from a wide range of stakeholders. 実践規範のプロセスと並行して、AI 事務局は、AI 法第 53 条(1)d) に基づき、汎用 AI モデルプロバイダが公開を義務付けられている、十分に詳細なトレーニングデータの要約に関するテンプレートも作成している。AI 事務局は、汎用 AI に関する幅広いマルチステークホルダー協議において、ステークホルダーから意見を集め、430 件以上の回答が幅広いステークホルダーから寄せられた。
The template for the summary of training data is closely linked to the providers' obligations in relation to transparency and copyright that have been detailed out in the Code of Practice. Given the close interlinkages of the two processes and the participation of all interested stakeholder groups with more than 1000 participants in the Code of Practice drafting, the AI Office has presented its preliminary ideas and allowed the participants to the Code to provide additional feedback on the preliminary structure and elements of the template. The topic is also discussed with the Member States' representatives in the AI Board subgroup and the European Parliament before the Commission adopts the template in the second quarter of 2025. トレーニングデータの要約テンプレートは、実践規範に詳述されている透明性および著作権に関するプロバイダの義務と密接に関連している。2 つのプロセスは密接に関連しており、実践規範の起草には 1,000 人以上の関係者グループが参加していることを踏まえ、AI 事務局は、その予備的な考え方を提示し、実践規範の参加者に対して、テンプレートの予備的な構造および要素について追加のフィードバックを提供する機会を設けた。また、2025 年第 2 四半期に委員会がテンプレートを採択する前に、AI 理事会小委員会および欧州議会の加盟国代表者ともこの問題について協議する予定である。

 

 


 

汎用AI実践規範への署名

・2025.07.17 Signing the General-Purpose AI code of practice

Signing the General-Purpose AI code of practice 汎用 AI 実践規範への署名
Questions and answers on signing the General-Purpose AI (GPAI) code of practice. 汎用 AI (GPAI) 実践規範への署名に関する質問と回答。
General FAQ 一般的な FAQ
Who can sign the code? 誰が規範に署名できるのか?
Any provider (or potential future provider) of a general-purpose AI model can sign the code by completing the Signatory Form and sending it to [mail] 汎用 AI モデルのあらゆるプロバイダ(または将来プロバイダとなる可能性のある者)は、署名者フォームに記入し、[mail]
に送信することで、規範に署名することができる。
The form should be signed by a person with sufficient authority to bind the provider to the General-Purpose code of Practice (e.g. a senior executive). このフォームには、プロバイダを汎用実践規範に拘束する十分な権限を有する者(上級幹部など)が署名する必要がある。
What does signing the code imply in terms of compliance with the AI Act? この規範に署名することは、AI 法への遵守に関してどのような意味があるか?
Once the code is assessed as adequate by the AI Office and the AI Board, providers of general-purpose AI models may rely on the code to demonstrate compliance with Articles 53 and 55 of the AI Act. Any opt-out from chapters of the code of practice results in losing the benefits of facilitating the demonstration of compliance in that respect. AI 事務局および AI 委員会によりこの規範が適切であると評価されると、汎用 AI モデルのプロバイダは、この規範を AI 法第 53 条および第 55 条への遵守の証明として利用することができる。実践規範の章の一部をオプトアウトすると、その部分に関する遵守の証明が容易になるというメリットは失われる。
For providers of general-purpose AI models that adhere to the code, the Commission will focus its enforcement activities on monitoring their adherence to the code. Because such providers are transparent about the measures they implement to comply with the AI Act, they benefit from increased trust from the Commission and other stakeholders. 実践規範を遵守する汎用 AI モデルのプロバイダについては、委員会は、実践規範の遵守状況を監視することに重点を置いて執行活動を行う。このようなプロバイダは、AI 法を遵守するために実施する措置について透明性を確保しているため、委員会やその他の利害関係者からの信頼を高めることができる。
What is the deadline for signing? 署名期限はいつか?
There is no deadline for signing the code. Providers may sign at any time. However, we ask existing providers to sign the code before August 1. For those signatories that agree, signatures will then be listed publicly online. 実践規範の署名期限はない。プロバイダはいつでも署名することができます。ただし、既存のプロバイダには 8 月 1 日までに署名をお願いしている。同意した署名者は、その署名がオンラインで公開される。
This way, the AI Office is aware of who intends to adhere to the code before the obligations of the AI Act start to apply on 2 August 2025. If a general-purpose AI model provider does not choose to adhere to the code, they need to demonstrate compliance through other means. これにより、AI 事務局は、2025 年 8 月 2 日に AI 法の義務が適用される前に、実践規範の遵守を表明している者を把握することができる。汎用 AI モデルプロバイダが実践規範の遵守を選択しなかった場合は、他の手段で遵守を実証する必要がある。
Can providers of general-purpose AI models without systemic risk sign up to the Safety and Security chapter? システミックリスクのない汎用 AI モデルのプロバイダは、「安全とセキュリティ」の章に署名することができるか?
Providers that are neither in the process of developing a general-purpose AI model with systemic risk nor intend to do so, may optionally sign the Safety and Security chapter. In this case, they are expected to specify which of their models - those that do not pose systemic risk - will be subject to the measures outlined in the Safety and Security chapter. システミックリスクのある汎用 AI モデルを開発中ではなく、その開発も予定していないプロバイダは、「安全とセキュリティ」の章に任意で署名することができる。この場合、システミックリスクのないモデルのうち、安全とセキュリティの章で規定されている措置の対象となるモデルを明記することが求められる。
Do signatories need to comply with the full code immediately? 署名者は、実践規範の全文を直ちに遵守する必要があるか?
The AI Office acknowledges that signatories may require some time to fully implement the measures in the code (for example, time required to implement the cybersecurity measures under Commitment 6 of the Safety and Security chapter), in light of the fact that the code was published only recently on July 10, 2025. If Signatories do not fully implement all commitments immediately after signing the code, the AI Office will not necessarily consider them to have broken their commitments under the code. Instead, in such cases, the AI Office will consider them to act in good faith and will be ready to collaborate to help ensure adherence to the code. AI オフィスは、実践規範が 2025 年 7 月 10 日に公表されたばかりであることを踏まえ、署名者が実践規範の措置を完全に実施するには、ある程度の時間が必要であることを認識している(例えば、安全・セキュリティ章のコミットメント 6 に定めるサイバーセキュリティ対策の実施に必要な時間)。署名者が実践規範署名後すぐにすべてのコミットメントを完全に実施しない場合、AI事務局は必ずしもそれらを実践規範上のコミットメント違反とみなさない。代わりに、そのような場合、AI事務局はそれらを善意に基づく行動とみなすとともに、実践規範遵守を確保するための協力を準備します。
What happens to providers that don’t adhere to the Code? 実践規範を遵守しないプロバイダはどうなるのか?
Providers that do not adhere to the code must demonstrate compliance via other adequate means and will have to report the measures they have implemented to the AI Office. Furthermore, such providers are expected to explain how the measures they implement ensure compliance with their obligations under the AI Act, for instance by carrying out a gap analysis that compares the measures they have implemented with the measures set out by the code. この規範を遵守しないプロバイダは、他の適切な手段で遵守を証明し、実施した措置を AI 事務局に報告しなければならない。さらに、そのようなプロバイダは、実施した措置が AI 法に基づく義務をどのように確保しているかを、例えば、実施した措置と規範で定められた措置を比較するギャップ分析を行うなどして説明することが求められる。
Providers that don’t adhere to the code may also be subject to a larger number of requests for information and requests for access to conduct model evaluations throughout the entire model lifecycle because the AI Office will have less of an understanding of how they are ensuring compliance with their obligations under the AI Act and will typically need more detailed information, including about modifications made to general-purpose AI models throughout their entire lifecycle. まこの規範を遵守しないプロバイダは、AI オフィスが AI 法に基づく義務の遵守状況を把握しにくいため、モデルライフサイクル全体を通じて、情報開示やモデル評価のためのアクセス要求がより多く発生する場合がある。また、AI オフィスは、モデルライフサイクル全体を通じて、汎用 AI モデルに加え、モデルライフサイクル全体における変更内容など、より詳細な情報を必要とする場合が多い。
How will the Code be updated? 実践規範はどのように更新されるのか?
In line with the Article 56(8) AI Act, the AI Office shall, as appropriate, encourage and facilitate the review and adaptation of the code. AI 法第 56 条 (8) に基づき、AI 事務局は、必要に応じて、規範の見直しおよび改訂を奨励し、促進するものとする。
The AI Office may facilitate formal updates to the code in response to technological developments, changes in the risk landscape, or experience with the application of the AI Act rules. The AI Office will review the code at least every two years, and it may propose a streamlined process for reviews and updates as needed. The AI Office may also issue guidance to further clarify concepts relevant for the code as appropriate to ensure its contribution to the proper application of the obligations under the AI Act. AI 事務局は、技術の発展、リスクの状況の変化、または AI 法規則の適用経験に応じて、規範の正式な更新を促進することができる。AI 事務局は、少なくとも 2 年に 1 回、規範を見直し、必要に応じて、見直しおよび更新のための簡素化されたプロセスを提案することができる。AI事務局は、AI法の義務の適切な適用への貢献を確保するため、必要に応じて、実践規範に関連する概念をさらに明確にするための指針を発行することができる。
Any updates to the text of the GPAI Code of Practice will be shared with the signatories for their consideration, and signatories will be invited to sign the updated code. GPAI実践規範の文言の更新は、署名者に共有され、検討が求められ、署名者は更新された実践規範に署名するよう招待される。
Can a signatory withdraw their signature? 署名者は署名を取り消すことができるか?
Yes, a signatory may withdraw their signature to the Code at any time. はい、署名者はいつでも実践規範への署名を撤回することができる。

 

 


 

AI 法における汎用 AI モデル – 質問と回答

・2025.07.16 General-Purpose AI Models in the AI Act – Questions & Answers

 

General-Purpose AI Models in the AI Act – Questions & Answers AI 法における汎用 AI モデル – 質問と回答
With this Q&A, the AI Office seeks to facilitate the interpretation of certain provisions of the AI Act. This Q&A does not constitute an official position of the Commission, and it is without prejudice to any decision or position of the Commission. Only the Court of Justice of the European Union is competent to authoritatively interpret the AI Act. この Q&A は、AI 法の一部の規定の解釈を容易にするために AI 事務局が作成したものです。この Q&A は、欧州委員会の公式見解を構成するものではなく、欧州委員会の決定や立場を妨げるものではない。AI 法を正式に解釈する権限は、欧州連合司法裁判所のみが有している。
General questions about the code of practice 実践規範に関する一般的な質問
Why do we need rules for general-purpose AI models? 汎用 AI モデルに関する規則が必要な理由
Artificial Intelligence (“AI”) promises huge benefits to our economy and society. General-purpose AI models play an important role in that regard, as they can be used for a variety of tasks and therefore form the basis for a range of downstream AI systems, used in Europe and worldwide. 人工知能(AI)は、経済および社会に大きなメリットをもたらすと期待されている。汎用 AI モデルは、さまざまなタスクに使用できるため、欧州および世界中で使用されるさまざまな下流 AI システムの基盤となる、その点で重要な役割を果たしている。
The AI Act aims to ensure that general-purpose AI models are safe and trustworthy. AI 法は、汎用 AI モデルの安全性と信頼性を確保することを目的としている。
To achieve that aim, it is crucial that providers of general-purpose AI models ensure a good understanding of their models along the entire AI value chain, so as to allow downstream providers both to integrate such models into AI systems and to fulfil their own obligations under the AI Act. More specifically, and as explained in more detail below, providers of general-purpose AI models must draw up technical documentation of their models to make available to downstream providers and to provide upon request to the AI Office and national competent authorities, put in place a copyright policy, and publish a training content summary. In addition, providers of general-purpose AI models posing systemic risks, which may be the case either because the models are very capable or because they have a significant impact on the internal market for other reasons, must notify those models to the Commission, assess and mitigate systemic risks stemming from those models, including via performing model evaluations, reporting serious incidents, and ensuring adequate cybersecurity of these models. その目的を達成するためには、汎用 AI モデルプロバイダが、AI バリューチェーン全体を通じて自社のモデルを十分に理解し、下流のプロバイダが AI システムに当該モデルを統合し、AI 法に基づく自らの義務を履行できるようにすることが重要である。より具体的には、以下で詳しく説明するように、汎用 AI モデルのプロバイダは、下流のプロバイダが利用できるように、また AI 事務局および各国の管轄当局の要請に応じて提供できるように、自社のモデルの技術文書を作成し、著作権ポリシーを策定し、トレーニングコンテンツの概要を公開しなければならない。 さらに、その能力が非常に高い、あるいはその他の理由により域内市場に重大な影響を及ぼすため、システミックリスクをもたらす可能性のある汎用 AI モデルのプロバイダは、そのモデルを欧州委員会に通知し、モデル評価の実施、重大なインシデントの報告、モデルの適切なサイバーセキュリティの確保などを通じて、そのモデルに起因するシステミックリスクを評価し、緩和しなければならない。
In this way, the AI Act contributes to safe and trustworthy innovation in the European Union.  このようにして、AI法は欧州連合における安全で信頼できるイノベーションに貢献する。
What are general-purpose AI models? 汎用AIモデルとは何か?
The AI Act defines a general-purpose AI model as “an AI model, including where such an AI model is trained with a large amount of data using self-supervision at scale, that displays significant generality and is capable of competently performing a wide range of distinct tasks regardless of the way the model is placed on the market and that can be integrated into a variety of downstream systems or applications” (Article 3(63) AI Act). AI 法では、汎用 AI モデルを「大規模な自己監督を用いて大量のデータで訓練された AI モデルを含め、市場での上市方法に関係なく、幅広い異なるタスクを適切に実行することができ、さまざまな下流のシステムやアプリケーションに統合できる、著しい汎用性を示す AI モデル」と定義している(AI 法第 3 条(63))。
The recitals to the AI Act further clarify which models should be deemed to display significant generality and to be capable of performing a wide range of distinct tasks. AI 法の前文では、どのモデルが「高い汎用性を示し、幅広い異なるタスクを遂行することができる」とみなされるべきかをさらに明確にしている。
According to recital 98 AI Act, “whereas the generality of a model could, inter alia, also be determined by a number of parameters, models with at least a billion of parameters and trained with a large amount of data using self-supervision at scale should be considered to display significant generality and to competently perform a wide range of distinctive tasks.” AI 法前文 98 項では、「モデルの汎用性は、とりわけ、パラメータの数によっても決定される可能性があるが、少なくとも 10 億個のパラメータを有し、大規模な自己監督を用いて大量のデータで訓練されたモデルは、高い汎用性を有し、幅広い多様なタスクを適切に実行できるとみなすべきである」と規定されている。
Recital 99 AI Act adds that “large generative AI models are a typical example for a general-purpose AI model, given that they allow for flexible generation of content, such as in the form of text, audio, images or video, that can readily accommodate a wide range of distinctive tasks.” AI 法前文 99 では、「大規模な生成的 AI モデルは、テキスト、音声、画像、動画などのコンテンツを柔軟に生成でき、幅広い特徴的なタスクに容易に対応できることから、汎用 AI モデルの典型的な例である」と付け加えられている。
Note that significant generality and ability to competently perform a wide range of distinctive tasks may be achieved by models within a single modality, such as text, audio, images, or video, if the modality is flexible enough. This may also be achieved by models that were developed, fine-tuned, or otherwise modified to be particularly good at a specific task or at a number of tasks in a specific domain. なお、そのモダリティが十分に柔軟であれば、テキスト、音声、画像、動画などの単一のモダリティのモデルでも、かなりの汎用性と幅広い特徴的なタスクを適切に実行できる能力を実現できることに留意してください。これは、特定のタスクまたは特定の分野における複数のタスクに特に優れるように開発、微調整、またはその他の方法で変更されたモデルでも実現できる。
Different stages of the development of a model do not constitute different models. モデルの開発の段階は、異なるモデルを構成するものではない。
The AI Office intends to provide further clarifications on what should be considered a general-purpose AI model, drawing on insights from the Commission’s Joint Research Centre, which is currently working on a scientific research project addressing this and other questions. AI 事務局は、この問題やその他の問題に取り組む科学研究プロジェクトを現在進行中の欧州委員会の共同研究センターからの知見を参考にし、汎用 AI モデルとみなすべきものについてさらに明確化を図る予定である。
What are general-purpose AI models with systemic risk? システミックリスクを伴う汎用 AI モデルとは何か?
Systemic risks are risks of large-scale harm from the most advanced (i.e. state-of-the-art) models at any given point in time or from other models that have an equivalent impact (see Article 3(65) AI Act). Such risks can manifest themselves, for example, through the lowering of barriers for chemical or biological weapons development, or unintended issues of control over autonomous general-purpose AI models (recital 110 AI Act). The most advanced models at any given point in time may pose systemic risks, including novel risks, as they are pushing the state of the art. At the same time, some models below the threshold reflecting the state of the art may also pose systemic risks, for example through reach, scalability, or scaffolding. システミックリスクとは、ある時点における最先端(すなわち、最新鋭)のモデル、または同等の影響力を持つ他のモデルによって生じる大規模な被害のリスクを指します(AI 法第 3 条(65)参照)。このようなリスクは、例えば、化学兵器や生物兵器の開発の障壁の低下、または自律型汎用 AI モデルの制御に関する意図しない問題として現れる可能性がある(AI 法前文 110)。ある時点における最先端のモデルは、その最先端の技術をさらに押し進めているため、新たなリスクを含むシステミックリスクをもたらす可能性がある。同時に、最先端の技術を反映した基準値を下回る一部のモデルも、その到達範囲、拡張性、または足場構築力などを通じて、システミックリスクをもたらす可能性がある。
Accordingly, the AI Act classifies a general-purpose AI model as a general-purpose AI model with systemic risk if it is one of the most advanced models at that point in time or if it has an equivalent impact (Article 51(1) AI Act). Which models are considered general-purpose AI models with systemic risk may change over time, reflecting the evolving state of the art and potential societal adaptation to increasingly advanced models. Currently, general-purpose AI models with systemic risk are developed by a handful of companies, although this may change in the future. したがって、AI 法では、その時点で最も先進的なモデルの一つである、あるいは同等の影響力を持つ汎用 AI モデルを、システムリスクのある汎用 AI モデルとして分類している(AI 法第 51 条第 1 項)。システムリスクのある汎用 AI モデルとみなされるモデルは、技術の進歩や、ますます高度化するモデルに対する社会の適応の程度に応じて、時間の経過とともに変化する可能性がある。現在、システムリスクのある汎用 AI モデルは、ごく一部の企業によって開発されているが、将来は変化する可能性がある。
To capture the most advanced models, i.e. models that match or exceed the capabilities recorded in the most advanced models so far, the AI Act lays down a threshold of 10^25 floating-point operations (FLOP) used for training the model (Article 51(1)(a) and (2) AI Act). Training a model that meets this threshold is currently estimated to cost tens of millions of Euros (Epoch AI, 2024). The AI Office is continuously monitoring technological and industrial developments, and the Commission may update the threshold, through the adoption of a delegated act (Article 51(3) AI Act), to ensure that it continues to single out the most advanced models as the state of the art evolves. For example, the value of the threshold itself could be adjusted and/or additional thresholds introduced. 最も高度なモデル、すなわちこれまで記録された最も高度なモデルと同等またはそれ以上の能力を有するモデルを捕捉するため、AI法は、モデルの訓練に用いられる浮動小数点演算(FLOP)の閾値を10^25と定めている(AI法第51条第1項(a)および(2))。この閾値を満たすモデルの訓練には、現在、数千万ユーロのコストがかかるものと推定されている(Epoch AI、2024年)。AI事務局は、技術的・産業的な動向を継続的に監視しており、委員会は、委任行為の採択(AI法第51条第3項)を通じて、閾値を更新し、最先端技術が進化するにつれて、最も高度なモデルを最先端技術として特定し続けるよう確保することができる。例えば、閾値自体の値を調整したり、追加の閾値を導入したりすることが考えられる。
To capture models with an impact equivalent to the most advanced models, the AI Act empowers the Commission to designate additional models as posing systemic risk, based on criteria such as the evaluations of capabilities of the model, number of users, scalability, or access to tools (Article 51(1)(b) and Annex XIII AI Act). 最先端のモデルと同等の影響力を持つモデルを捕捉するため、AI 法は、モデルの機能の評価、ユーザー数、拡張性、ツールへのアクセスなどの規準に基づき、システムリスクをもたらす追加のモデルを指定する権限を委員会に与えている(AI 法第 51 条(1)(b) および附属書 XIII)。
The AI Office intends to provide further clarifications on how general-purpose AI models will be classified as general-purpose AI models with systemic risk, drawing on insights from the Commission’s Joint Research Centre which is currently working on a scientific research project addressing this and other questions. AI 事務局は、この問題やその他の問題に取り組む科学研究プロジェクトを現在進行中の欧州委員会の共同研究センターからの知見を参考にし、汎用 AI モデルが、システムリスクのある汎用 AI モデルとしてどのように分類されるかについて、さらに明確化を行う予定である。
What is a provider of a general-purpose AI model? 汎用 AI モデルのプロバイダとは何か?
The AI Act rules on general-purpose AI models apply to providers placing such models on the market in the Union, irrespective of whether those providers are established or located within the Union or in a third country (Article 2(1)(a) AI Act). 汎用 AI モデルに関する AI 法の規定は、そのプロバイダが EU 域内に設立または所在するか、あるいは第三国に設立または所在するかに関わらず、EU 域内でそのようなモデルを上市するプロバイダに適用されます(AI 法第 2 条(1)(a))。
A provider of a general-purpose AI model means a natural or legal person, public authority, agency or other body that develops a general-purpose AI model or that has such a model developed and places it on the market, whether for payment or free or charge (Article 3(3) AI Act). 汎用 AI モデルのプロバイダとは、汎用 AI モデルを開発する、またはその開発を委託し、有償または無償で上市する自然人、法人、公的機関、機関、その他の団体をいう(AI 法第 3 条(3))。
To place a model on the market means the first making available of the model on the Union market (Article 3(9) AI Act), that is, to supply it for distribution or use on the Union market for the first time in the course of a commercial activity, whether in return for payment or free of charge (Article 3(10) AI Act). Note that a general-purpose AI model is also considered to be placed on the market if that model’s provider integrates the model into its own AI system which is made available on the market or put into service, unless the model is (a) used for purely internal processes that are not essential for providing a product or a service to third parties, (b) the rights of natural persons are not affected, and (c) the model is not a general-purpose AI model with systemic risk (recital 97 AI Act). モデルを市場に上市とは、そのモデルを EU 市場で初めて入手可能にすることをいう(AI 法第 3 条(9))。つまり、有償または無償を問わず、商業活動において EU 市場で流通または使用するために、そのモデルを初めて供給することをいう(AI 法第 3 条(10))。なお、汎用 AI モデルは、そのモデルの提供者が、市場で入手可能または使用可能な自社の AI システムに当該モデルを統合した場合も、上市されたものとみなされます。ただし、以下の条件を満たす場合を除きます。(a) サードパーティに製品またはサービスを提供するために不可欠ではない、純粋に内部的なプロセスに使用されている場合 (b) 自然人の権利が影響を受けない場合、および(c) モデルが、システムリスクを伴う汎用 AI モデルではない場合(AI 法前文 97)。
What are the obligations for providers of general-purpose AI models? 汎用 AI モデルのプロバイダにはどのような義務があるか?
Providers of general-purpose AI models must document technical information about their models for the purpose of providing that information upon request to the AI Office and national competent authorities (Article 53(1)(a) AI Act) and making it available to downstream providers (Article 53(1)(b) AI Act). They must also put in place a policy to comply with Union law on copyright and related rights (Article 53(1)(c) AI Act) and draw up and make publicly available a sufficiently detailed summary about the content used for training the model (Article 53(1)(d) AI Act). 汎用 AI モデルのプロバイダは、AI 事務局および各国の管轄当局からの要請に応じてその情報を提供するため(AI 法第 53 条(1)(a))、また下流のプロバイダが利用できるようにするため(AI 法第 53 条(1)(b))、そのモデルに関する技術情報を文書化しなければならない。また、著作権および関連する権利に関する連合法令に準拠するためのポリシーを策定し(AI法53条1項(c))、モデルの訓練に使用されたコンテンツに関する十分な詳細な要約を作成し、公開しなければならない(AI法53条1項(d))。
The General-Purpose AI Code of Practice should provide signatories with further detail on how to ensure compliance with these obligations in the sections dealing with transparency and copyright (led by Working Group 1). 汎用 AI 実践規範は、透明性および著作権に関するセクション(ワーキンググループ 1 が主導)において、これらの義務の遵守を確保する方法について、署名者にさらに詳細な情報を提供すべきである。
Under the AI Act, providers of general-purpose AI models with systemic risk have additional obligations. They must assess and mitigate systemic risks, in particular by performing model evaluations, keeping track of, documenting, and reporting serious incidents, and ensuring adequate cybersecurity protection for the model and its physical infrastructure (Article 55 AI Act). AI 法に基づき、システミックリスクのある汎用 AI モデルのプロバイダには追加の義務が課せられる。彼らは、特に、モデル評価の実施、重大なインシデントの追跡、文書化、報告、およびモデルとその物理的インフラストラクチャの適切なサイバーセキュリティ保護の確保により、システミックリスクを評価し、緩和しなければならない(AI 法第 55 条)。
The General-Purpose AI Code of Practice should provide signatories with further detail on how to ensure compliance with these obligations in the sections dealing with systemic risk assessment, technical risk mitigation, and governance risk mitigation (led by Working Groups 2, 3, and 4 respectively). 汎用 AI 実践規範は、システミックリスクの評価、技術的リスクの緩和、およびガバナンスリスクの緩和に関するセクション(それぞれワーキンググループ 2、3、4 が主導)において、これらの義務の遵守を確保する方法について、署名者にさらなる詳細情報を提供すべきである。
The obligations for providers of general-purpose AI models apply from 2 August 2025 (Article 113(b) AI Act), with special rules for general-purpose AI models placed on the market before that date (Article 111(3)) AI Act. 汎用 AI モデルのプロバイダに対する義務は、2025 年 8 月 2 日から適用される(AI 法第 113 条(b))。ただし、その日付より前に上市された汎用 AI モデルについては、特別規則が適用される(AI 法第 111 条(3))。
What if someone open-sources a model, do they have to comply with the obligations for providers of general-purpose AI models? 誰かがモデルをオープンソース化した場合、その人は汎用 AI モデルのプロバイダに対する義務を遵守しなければならないのか?
The obligations to draw up and provide documentation to the AI Office, national competent authorities, and downstream providers (Article 53(1)(a) and (b) AI Act) do not apply if the model is released under a free and open-source license and its parameters, including the weights, the information on the model architecture, and the information on model usage, are made publicly available (Article 53(2) AI Act). Recitals 102 and 103 AI Act further clarify what constitutes a free and open-source license and the AI Office intends to provide further clarifications on questions concerning open-sourcing general-purpose AI models. AI 事務局、国の管轄当局、および下流プロバイダに対して文書を作成し、提供するという義務(AI 法第 53 条(1)(a)および(b))は、モデルがフリーでオープンソースのライセンスに基づいて公開され、そのパラメータ(重み、モデルアーキテクチャに関する情報、モデル使用に関する情報など)が一般に公開されている場合には適用されない(AI 法第 53 条(2))。AI 法前文 102 項および 103 項では、フリーでオープンソースのライセンスとは何であるかをさらに明確化しており、AI 事務局は、汎用 AI モデルのオープンソース化に関する質問について、さらに明確化を図る予定である。
The exemption does not apply to general-purpose AI models with systemic risk. After the open-source model release, measures necessary to ensure compliance with the obligations of Article 55 may be more difficult to implement (Recital 112 AI Act). Consequently, providers of general-purpose AI models with systemic risk may need to assess and mitigate systemic risks before releasing their models as open-source. この適用除外は、システミックリスクのある汎用 AI モデルには適用されない。オープンソースモデルがリリースされた後、第 55 条の義務を確実に遵守するために必要な措置の実施がより困難になる可能性がある(AI 法前文 112)。したがって、システミックリスクのある汎用 AI モデルのプロバイダは、モデルをオープンソースとしてリリースする前に、システミックリスクを評価し、緩和する必要があるかもしれない。
The General-Purpose AI Code of Practice should provide signatories with further detail on how to comply with the obligations in Articles 53 and 55 in relation to the different ways of releasing general-purpose AI models, including open-sourcing. 汎用 AI 実践規範は、オープンソース化を含む汎用 AI モデルのさまざまな公開方法に関して、第 53 条および第 55 条に定める義務を遵守する方法について、署名者にさらに詳細な情報を提供すべきである。
An important but difficult question underpinning the process of drawing up the Code of Practice is that of finding a balance between pursuing the benefits and mitigating the risks from the open-sourcing of advanced general-purpose AI models: open-sourcing advanced general-purpose AI models may indeed yield significant societal benefits, including through fostering AI safety research; at the same time, when such models are open-sourced, risk mitigations are more easily circumvented or removed. 実践規範の策定プロセスにおける重要かつ困難な課題は、高度な汎用 AI モデルのオープンソース化によるメリットの追求とリスクの緩和とのバランスを見出すことである。高度な汎用 AI モデルをオープンソース化することで、AI の安全性に関する研究の推進など、社会的に大きなメリットが得られる可能性がある一方で、こうしたモデルがオープンソース化されると、リスクの緩和策が回避されたり、無効化されたりする可能性が高まる。
Do the obligations for providers of general-purpose AI models apply in the development phase? 汎用 AI モデルのプロバイダに対する義務は、開発段階にも適用されるのか?
Article 2(8) AI Act provides that, as a general matter, the AI Act “does not apply to any research, testing or development activity regarding AI systems or AI models prior to their being placed on the market or put into service.” AI 法第 2 条(8) は、AI 法は、原則として、「AI システムまたは AI モデルが上市または使用開始される前の、AI システムまたは AI モデルに関する研究、試験、開発活動には適用されない」と規定している。
At the same time, certain obligations for providers of general-purpose AI models (with and without systemic risk) explicitly or implicitly pertain to the development phase of models intended for, but prior to, the placing on the market. This is the case, for example, for the obligations for providers to notify the Commission that their general-purpose AI model meets or will meet the training compute threshold (Articles 51 and 52 AI Act), to document information about training and testing (Article 53 AI Act), and to assess and mitigate systemic risk (Article 55 AI Act). In particular, Article 55(1)(b) AI Act explicitly provides that “providers of general-purpose AI models with systemic risk shall assess and mitigate possible systemic risks at Union level, including their sources, that may stem from the development (...) of general-purpose AI models with systemic risk. 同時に、汎用 AI モデル(システミックリスクの有無を問わず)のプロバイダに対する一定の義務は、上市前のモデルの開発段階にも明示的または黙示的に適用される。例えば、プロバイダは、汎用 AI モデルがトレーニング計算のしきい値を満たしているか、または満たす予定であることを委員会に通知する義務(AI 法第 51 条および第 52 条)、トレーニングおよびテストに関する情報を文書化する義務(AI 法第 53 条)、およびシステミックリスクを評価し、緩和する義務(AI 法第 55 条)がある。特に、AI 法第 55 条 (1) (b) は、「システミックリスクのある汎用 AI モデルのプロバイダは、システミックリスクのある汎用 AI モデルの開発から生じる可能性のある、その発生源を含む、EU レベルでのシステミックリスクを評価し、緩和しなければならない」と明示的に規定している。
The AI Office expects discussions with providers of general-purpose AI models with systemic risk to start early in the development phase. This is consistent with the obligation for providers of general-purpose AI models that meet the training compute threshold laid down in Article 51(2) AI Act to “notify the Commission without delay and in any event within two weeks after that requirement is met or it becomes known that it will be met” (Article 52(1) AI Act). Indeed, training of general-purpose AI models takes considerable planning, which includes the upfront allocation of compute resources, and providers of general-purpose AI models are therefore able to know if their model will meet the training compute threshold before the training is complete (recital 112 AI Act). AI 事務局は、開発段階の早い段階で、システムリスクのある汎用 AI モデルのプロバイダとの協議を開始することを期待している。これは、AI 法第 51 条(2)で規定されているトレーニング計算のしきい値を満たす汎用 AI モデルのプロバイダが、「その要件を満たした、またはその要件を満たすことが判明した場合は、遅滞なく、かつ、いかなる場合でもその要件を満たしてから 2 週間以内に、欧州委員会に通知する」という義務と一致している(AI 法第 52 条(1))。実際、汎用 AI モデルのトレーニングには、計算リソースの事前割り当てなど、かなりの計画が必要であり、汎用 AI モデルのプロバイダは、トレーニングが完了する前に、そのモデルがトレーニング計算のしきい値を満たすかどうかを知ることができる(AI 法前文 112)。
The AI Office intends to provide further clarifications on this question. AI 事務局は、この問題についてさらに明確化を行う予定である。
If someone fine-tunes or otherwise modifies a model, do they have to comply with the obligations for providers of general-purpose AI models? モデルを微調整またはその他の方法で変更した場合、汎用 AI モデルのプロバイダとしての義務を遵守しなければならないか?
General-purpose AI models may be further modified or fine-tuned into new models (recital 97 AI Act). Accordingly, downstream entities that fine-tune or otherwise modify an existing general-purpose AI model may become providers of new models. The specific circumstances in which a downstream entity becomes a provider of a new model is a difficult question with potentially large economic implications, since many organisations and individuals fine-tune or otherwise modify general-purpose AI models developed by another entity. 汎用 AI モデルは、さらに変更または微調整して新しいモデルにすることができる(AI 法前文 97)。したがって、既存の汎用 AI モデルを微調整またはその他の方法で変更する下流事業体は、新しいモデルのプロバイダとなる可能性がある。下流事業体が新しいモデルのプロバイダとなる具体的な状況は、多くの組織や個人が他の事業体が開発した汎用 AI モデルを微調整またはその他の方法で変更しているため、経済的に大きな影響をもたらす可能性のある難しい問題である。
In the case of a modification or fine-tuning of an existing general-purpose AI model, the obligations for providers of general-purpose AI models in Article 53 AI Act should be limited to the modification or fine-tuning, for example, by complementing the already existing technical documentation with information on the modifications (Recital 109). The obligations for providers of general-purpose AI models with systemic risk in Article 55 AI Act should only apply in clearly specified cases. The AI Office intends to provide further clarifications on this question. 既存の汎用 AI モデルを変更または微調整する場合、AI 法第 53 条における汎用 AI モデルのプロバイダの義務は、例えば、既存の技術文書に変更に関する情報を追加するなど、その変更または微調整に限定すべきである。(前文 109)。AI 法第 55 条における、システミックリスクを伴う汎用 AI モデルのプロバイダの義務は、明確に規定された場合にのみ適用されるべきである。AI 事務局は、この問題についてさらに明確化を図る予定である。
Regardless of whether a downstream entity that incorporates a general-purpose AI model into an AI system is deemed to be a provider of the general-purpose AI model, that entity must comply with the relevant AI Act requirements and obligations for AI systems. 汎用 AI モデルを AI システムに組み込む下流の事業体が、汎用 AI モデルのプロバイダとみなされるかどうかに関わらず、その事業体は、AI 法における AI システムに関する要件および義務を遵守しなければならない。
What is the General-Purpose AI Code of Practice? 汎用 AI 実践規範とは何か?
Based on Article 56 of the AI Act, the General-Purpose AI Code of Practice should detail the manner in which providers of general-purpose AI models and of general-purpose AI models with systemic risk may comply with their obligations under the AI Act. The AI Office is facilitating the drawing-up of this Code of Practice, with four working groups chaired by independent experts and involving nearly 1000 stakeholders, EU Member States representatives, as well as European and international observers. AI 法第 56 条に基づき、汎用 AI 実践規範は、汎用 AI モデルおよびシステムリスクを伴う汎用 AI モデルのプロバイダが AI 法に基づく義務を遵守する方法を詳細に規定するものとする。AI 事務局は、独立した専門家が議長を務める 4 つの作業部会と、1000 人近くの利害関係者、EU 加盟国の代表者、欧州および国際的なオブザーバーが参加して、この規範の策定を推進している。
More precisely, the Code of Practice should detail at least how providers of general-purpose AI models may comply with the obligations laid down in Articles 53 and 55 AI Act. This means that the Code of Practice can be expected to have two parts: one that applies to providers of all general-purpose AI models (Article 53), and one that applies only to providers of general-purpose AI models with systemic risk (Article 55). より正確には、実践規範には、汎用 AI モデルの提供者が AI 法第 53 条および第 55 条に規定される義務をどのように遵守すべきかを、少なくとも詳細に規定すべきである。つまり、実践規範は 2 つの部分で構成されることが予想されます。1 つは、すべての汎用 AI モデルのプロバイダに適用される部分(第 53 条)、もう 1 つは、システムリスクのある汎用 AI モデルのプロバイダにのみ適用される部分(第 55 条)です。
What is not part of the Code of Practice? 実践規範には何が含まれないのか?
The Code of Practice should not address inter alia the following issues: defining key concepts and definitions from the AI Act (such as “general-purpose AI model”), updating the criteria or thresholds for classifying a general-purpose AI model as a general-purpose AI model with systemic risk (Article 51), outlining how the AI Office will enforce the obligations for providers of general-purpose AI models (Chapter IX Section 5), and questions concerning fines, sanctions, and liability. 実践規範は、とりわけ以下の事項については扱わないものとする。AI 法における重要な概念や定義(「汎用 AI モデル」など)の定義、汎用 AI モデルをシステムリスクのある汎用 AI モデルとして分類するための規準や閾値の更新(第 51 条)、AI 事務局が汎用 AI モデルのプロバイダに対する義務をどのように執行するかについての概要(第 IX 章第 5 節)、罰金、制裁、責任に関する問題。
These issues may instead be addressed through other means (decisions, delegated acts, implementing acts, further communications from the AI Office, etc.). これらの事項は、他の手段(決定、委任行為、実施行為、AI 事務局からのさらなるコミュニケーションなど)によって対処することができる。
Nevertheless, the Code of Practice may include commitments by providers of general-purpose AI models that sign the Code to document and report additional information, as well as to involve the AI Office and third parties throughout the entire model lifecycle, in so far as this is considered necessary for providers to effectively comply with their obligations under the AI Act. ただし、実践規範には、AI 法に基づく義務を効果的に遵守するために必要と認められる限り、実践規範に署名する汎用 AI モデルの提供者が、追加情報を文書化し報告すること、およびモデルのライフサイクル全体を通じて AI 事務局およびサードパーティを関与させることを約束する条項を含めることができる。
Do AI systems play a role in the Code of Practice? AI システムは実践規範において何らかの役割を果たすのか?
The AI Act distinguishes between AI systems and AI models, imposing requirements for certain AI systems (Chapters II-IV) and obligations for providers of general-purpose AI models (Chapter V). While the provisions of the AI Act concerning AI systems depend on the context of use of the system, the provisions of the AI Act concerning general-purpose AI models apply to the model itself, regardless of what is or will be its ultimate use. The Code of Practice should only pertain to the obligations in the AI Act for providers of general-purpose AI models. AI 法は、AI システムと AI モデルを区別し、特定の AI システムに要件(第 II 章から第 IV 章)を課し、汎用 AI モデルのプロバイダに義務を課している (第 V 章)を規定している。AI 法における AI システムに関する規定は、システムの使用状況に応じて適用されるが、汎用 AI モデルに関する規定は、その最終的な使用目的にかかわらず、モデル自体に適用される。実践規範は、汎用 AI モデルのプロバイダに対する AI 法の義務のみに関係する。
Nevertheless, there are interactions between the two sets of rules, as general-purpose AI models are typically integrated into and form part of AI systems. If a provider of a general-purpose AI model integrates that model into an AI system, that provider must comply with the obligations for providers of general-purpose AI models and, if the AI system falls within the scope of the AI Act, with the requirements for AI systems. If a downstream provider integrates a general-purpose AI model into an AI system, the provider of the general-purpose AI model must cooperate with the downstream provider of the AI system to ensure that the latter can comply with its obligations under the AI Act if the AI system falls within the scope of the AI Act (for example by providing certain information to the downstream provider). ただし、汎用 AI モデルは通常 AI システムに統合され、その一部を構成するため、2 つの規則セット間には相互作用がある。汎用 AI モデルの提供者がそのモデルを AI システムに統合する場合、その提供者は、汎用 AI モデルの提供者に課せられる義務、および AI システムが AI 法の適用範囲に該当する場合には AI システムに関する要件を遵守しなければならない。下流のプロバイダが汎用 AI モデルを AI システムに統合する場合、 AI システムが AI 法の適用範囲に該当する場合には、汎用 AI モデルの提供者は、AI システムのダウンストリームプロバイダが AI 法に基づく義務を確実に遵守できるよう、AI システムのダウンストリームプロバイダと協力しなければならない (例えば、下流のプロバイダに特定の情報を提供するなど)。
Given these interactions between models and systems, and between the obligations and requirements for each, an important question underlying the Code of Practice concerns which measures are appropriate at the model layer, and which need to be taken at the system layer instead. モデルとシステム、およびそれぞれの義務と要件の間にこのような相互作用があることから、実践規範の根底にある重要な問題は、モデル層で適切な措置はどれか、そしてシステム層で講じるべき措置はどれかということだ。
How does the Code of Practice take into account the needs of start-ups? 実践規範は、スタートアップのニーズをどのように考慮しているか?
The Code of Practice should set out its objectives, commitments, measures and, as appropriate, key performance indicators (KPIs) to measure the achievement of its objectives. Commitments, measures and KPIs related to the obligations applicable to providers of all general-purpose AI models should take due account of the size of the provider and allow simplified ways of compliance for SMEs, including start-ups, that should not represent an excessive cost and not discourage the use of such models (Recital 109 AI Act). Moreover, the reporting commitments related to the obligations applicable to providers of general-purpose AI models with systemic risk should reflect differences in size and capacity between various providers (Article 56(5) AI Act), while ensuring that they are proportionate to the risks (Article 56(2)(d) AI Act). 実践規範は、その目的、コミットメント、措置、および必要に応じて、その目的達成度を測定するための主要業績評価指標(KPI)を定めるべきである。すべての汎用 AI モデルのプロバイダに適用される義務に関連するコミットメント、措置、および KPI は、プロバイダの規模を十分に考慮し、スタートアップを含む中小企業に対して、過度のコストを課すことなく、またそのようなモデルの使用を妨げるようなことのない、簡素化された遵守方法を認めるべきである(AI 法前文 109)。さらに、システミックリスクのある汎用 AI モデルのプロバイダに適用される義務に関する報告のコミットメントは、さまざまなプロバイダの規模や能力の違いを反映するとともに、リスクに見合ったものであることを確保すべきである(AI 法第 56 条(5))。
When will the Code of Practice be finalised? 実践規範はいつ確定するのか?
After the publication of the third draft of the Code of Practice, it is expected that there will be one more drafting round over the coming months. Thirteen Chairs and Vice-Chairs, drawn from diverse backgrounds in computer science, AI governance and law, are responsible for synthesizing submissions from a multi-stakeholder consultation and discussions with the Code of Practice Plenary consisting of around 1000 stakeholders. This iterative process will lead to a final Code of Practice which should reflect the various submissions whilst ensuring an appropriate reflection of how to comply with the AI Act. 実践規範の 3 番目のドラフトが公表された後、今後数カ月にわたってさらに 1 回、草案の策定作業が行われる予定である。コンピュータサイエンス、AI ガバナンス、法律などのさまざまな分野から選ばれた 13 人の議長および副議長が、多国間協議で寄せられた意見や、約 1000 人の利害関係者で構成される実践規範総会での議論をまとめます。この反復プロセスを経て、さまざまな意見が反映され、AI 法への遵守方法を適切に反映した最終的な実践規範が策定される予定である。
Based on Article 56(9), “codes of practice shall be ready at the latest by 2 May 2025” in order to give providers sufficient time before the GPAI rules become applicable on 2 August 2025. 第 56 条 (9) 項に基づき、「実践規範は、GPAI 規則が 2025 年 8 月 2 日に適用されるまでに、プロバイダに十分な時間を確保するため、2025 年 5 月 2 日までに策定される」ことになっている。
Once the Code is finalised “the AI Office and the Board shall assess whether the codes of practice cover the obligations provided for in Articles 53 and 55 (...). They shall publish their assessment of the adequacy of the codes of practice. The Commission may, by way of an implementing act, approve a code of practice and give it a general validity within the Union.” (Article 56(6) AI Act). 実践規範が確定すると、「AI 事務局および委員会は、実践規範が第 53 条および第 55 条で規定される義務を網羅しているかどうかを評価する。彼らは、実践規範の妥当性に関する評価を公表する。委員会は、実施法によって、実践規範を承認し、EU 域内でその一般的有効性を付与することができる」と規定されている(AI 法第 56 条(6))。
What are the legal effects of the Code of Practice? 実践規範の法的効力とは何か?
If approved via implementing act, the Code of Practice obtains general validity, meaning that adherence to the Code of Practice becomes a means to demonstrate compliance with the AI Act, while not providing a presumption of conformity with the AI Act. Providers may also demonstrate compliance with the AI Act in other ways. In that case, providers would be expected to adopt measures to ensure compliance with their AI Act obligations that are adequate, effective, and proportionate, and which may include reporting to the AI Office. 実施法令により承認された場合、実践規範は一般的な効力を有することになり、実践規範の遵守は AI 法への準拠を示す手段となるが、AI 法への準拠を推定するものではない。プロバイダは、他の方法でも AI 法への準拠を証明することができる。その場合、プロバイダは、AI 法に基づく義務を遵守するために、適切かつ効果的かつ均衡のとれた措置を講じることを期待される。その措置には、AI 事務局への報告も含まれる。
Based on the AI Act, additional legal effects of the Code of Practice are that the AI Office may take a provider’s adherence to the Code of Practice into account when monitoring its effective implementation and compliance with the AI Act (Article 89(1) AI Act) and may favourably take into account commitments made in the Code of Practice when fixing the amount of fines depending on the specific circumstances (Article 101(1) AI Act). AI 法に基づき、実践規範の追加的な法的効果としては、AI 事務局が、その効果的な実施および AI 法への遵守を監視する際に、プロバイダの実践規範の遵守を考慮することができる(AI 法第 89 条(1))、および特定の状況に応じて罰金の額を決定する際に、実践規範におけるコミットメントを有利に考慮することができる(AI 法第 101 条(1))ことが挙げられる。
Finally, “If, by 2 August 2025, a code of practice cannot be finalised, or if the AI Office deems it is not adequate following its assessment under paragraph 6 of this Article, the Commission may provide, by means of implementing acts, common rules for the implementation of the obligations provided for in Articles 53 and 55, including the issues set out in paragraph 2 of this Article.” (Article 56(9) AI Act). 最後に、「2025 年 8 月 2 日までに実践規範が確定できない場合、または AI 事務局が本条第 6 項に基づくアセスメントの結果、実践規範が適切でないと判断した場合、委員会は、実施法令により、本条第 2 項で規定する事項を含む、第 53 条および第 55 条に規定する義務の実施に関する共通規則を定めることができる」と規定されている。(AI 法第 56 条(9))。
How will the Code of Practice be reviewed and updated? 実践規範はどのように見直され、更新されるのか?
Based on Article 56(8) AI Act, “the AI Office shall, as appropriate, also encourage and facilitate the review and adaptation of the codes of practice, in particular in light of emerging standards”. The third draft of the Code of Practice includes a suggested mechanism for its review and updating. AI 法第 56 条(8)に基づき、「AI 事務局は、必要に応じて、特に新たな標準の出現を踏まえて、実践規範の見直しと適応を奨励し、促進するものとする」。実践規範の 3 番目のドラフトには、その見直しと更新のための仕組み案が含まれている。
Which enforcement powers does the AI Office have? AI 事務局にはどのような執行権限があるか?
The AI Office will supervise and enforce the obligations laid down in the AI Act for providers of general-purpose AI models (Article 88 AI Act) and, exceptionally, the obligations for providers of AI systems based on general purpose AI models if the provider of the model and the system are the same (Article 75, (1) AI Act). The AI Office will support the relevant market surveillance authorities of the Member States in their enforcement of the requirements for AI systems (Article 75 (2) and (3) AI Act), among other tasks. Enforcement by the AI Office is underpinned by the powers given to it by the AI Act, namely the powers to request information (Article 91 AI Act), conduct evaluations of general-purpose AI models (Article 92 AI Act), request measures from providers, including implementing risk mitigations, and recalling the model from the market (Article 93 AI Act), and to impose fines of up to 3% of global annual turnover or 15 million Euros, whichever is higher (Article 101 AI Act). AI 事務局は、汎用 AI モデルの提供者に AI 法で定められた義務の監督および執行を行う(AI 法第 88 条)。また、例外的に、モデルとシステムの提供者が同一である場合は、汎用 AI モデルに基づく AI システムの提供者に義務の監督および執行を行う(AI 法第 75 条(1))。AI 事務局は、加盟国の関連市場監視当局が AI システムに関する要件を執行する支援を行う(AI 法第 75 条(2)および(3))。AI 事務局による執行は、AI 法により AI 事務局に付与された権限、すなわち、情報要求権限 (AI 法第 91 条)、汎用 AI モデルの評価(AI 法第 92 条)、プロバイダに対するリスクの緩和措置の実施や市場からのモデルの回収などの措置の要請(AI 法第 93 条)、および世界全体の年間売上高の 3% または 1,500 万ユーロのうちいずれか高い方の罰金の賦課(AI 法第 101 条)である。

 

 


 

汎用Ai実践規範に関する追加情報

・2025.07.11 Questions and answers on the code of practice for General-Purpose AI

 

Questions and answers on the code of practice for General-Purpose AI 汎用AI実践規範に関する質問と回答
This Q&A gives additional information about the General-Purpose AI Code of Practice. このQ&Aは、汎用AI実践規範に関する追加情報を提供する。
General FAQ 一般的な FAQ
What is the General-Purpose AI Code of Practice? 汎用 AI 実践規範とは何か?
The General-Purpose AI Code of Practice is a voluntary tool, prepared by independent experts in a multi-stakeholder process, designed to help industry comply with the AI Act’s obligations for providers of general-purpose AI models, ensuring that general-purpose AI models placed on the European market are safe and transparent, including the most powerful ones. 汎用 AI 実践規範は、多角的な利害関係者によるプロセスを通じて、独立した専門家によって作成された自主的なツールであり、業界が汎用 AI モデルの提供者に課せられた AI 法上の義務を遵守し、欧州市場に上市される汎用 AI モデル(最も強力なモデルを含む)の安全性と透明性を確保することを目的としている。
It outlines a way for providers of general-purpose AI models and of general-purpose AI models with systemic risk to demonstrate compliance with the AI Act’s relevant obligations. この規範は、汎用 AI モデルおよびシステミックリスクのある汎用 AI モデルのプロバイダが、AI 法に関連する義務の遵守を実証する方法を概説している。
The Code has three chapters: Transparency and Copyright, both addressing all providers of general-purpose AI models, and Safety and Security, relevant only to a limited number of providers of the most advanced models, subject to the AI Act's obligations for providers of general-purpose AI models with systemic risk. この規範は 3 章で構成されており、「透明性」と「著作権」は汎用 AI モデルのプロバイダすべてを対象とし、「安全とセキュリティ」は、AI 法がシステミックリスクのある汎用 AI モデルのプロバイダに課す義務の対象となる、ごく一部の最先端モデルのプロバイダのみを対象としている。
The AI Act defines systemic risk as specific to high-impact capabilities, i.e. capabilities that match or exceed the capabilities of the most advanced general-purpose AI models, that have a significant impact on the Union market. The AI Act currently presumes that models trained with a cumulative amount of compute greater than 10^25 floating-point operations possess high-impact capabilities. AI 法では、システムリスクは、EU 市場に重大な影響を与える、最も先進的な汎用 AI モデルの能力と同等またはそれを上回る、影響力の大きい能力に特有のものとして定義されている。AI 法は現在、累積計算量が 10^25 倍以上の浮動小数点演算でトレーニングされたモデルは、影響力の大きい能力を有すると推定している。
What was the process of drafting the Code and who was involved? 実践規範のドラフトはどのようなプロセスで行われ、誰が関与したのか?
The Code was developed through an extensive multi-stakeholder process launched in July 2024, involving over 1 400 participants from industry, academia, civil society, rightsholders, and EU Member States represented in the AI Board. この規範は、2024年7月に開始された、業界、学界、市民社会、権利者、AI 委員会に代表される EU 加盟国など、1,400 人以上の参加者が関与した広範なマルチステークホルダープロセスを通じて作成されました。
Thirteen independent experts appointed by the AI Office led the drafting process. They developed the Code through three rounds of consultations, incorporating over 1 600 written submissions and feedback from 40 workshops. AI 事務局が任命した 13 人の独立専門家が、ドラフト作成プロセスを主導した。彼らは 3 回にわたる協議を経て、1,600 件以上の書面提出と 40 件のワークショップからのフィードバックを取り入れて、この規範を策定した。
How can general-purpose AI providers adhere to the Code, and what are the benefits? 汎用 AI プロバイダは、この規範をどのように遵守することができ、そのメリットは何か?
All providers of general-purpose AI models with existing or planned operations in the EU market will be invited to adhere to the Code, with the AI Office providing details on the process soon. EU 市場で事業を展開している、または事業展開を計画している汎用 AI モデルのすべてのプロバイダは、この規範の遵守を求められます。AI 事務局は、そのプロセスに関する詳細をまもなく発表する予定です。
Adhering to a Code assessed as adequate by the AI Office and the Board will offer a simple and transparent way to demonstrate compliance with the AI Act. This offers a streamlined compliance process, with enforcement focused on monitoring their adherence to the Code, resulting in greater predictability and reduced administrative burden. AI 事務局および委員会が適切と評価した規範を遵守することで、AI 法への準拠を簡単かつ透明性の高い方法で実証することができる。これにより、規範の遵守状況を監視することに重点を置いた執行により、コンプライアンスプロセスが効率化され、予測可能性が高まり、行政負担が軽減される。
Does the Code impose additional obligations beyond the AI Act? この規範は、AI 法に追加の義務を課すものなのか?
The Code does not impose obligations. It serves as guidance to help providers meet their existing obligations under the AI Act without creating new ones, extending existing ones, or imposing additional burdens. It serves as a voluntary tool that helps providers demonstrate compliance with the AI Act's binding provisions, without exceeding its scope or imposing additional burdens. この規範は義務を課すものではない。これは、プロバイダが AI 法に基づく既存の義務を履行するためのガイダンスとして機能し、新たな義務を課したり、既存の義務を拡大したり、追加的な負担を課したりするものではない。これは、AI 法の拘束力のある規定の範囲を超えたり、追加的な負担を課したりすることなく、プロバイダが AI 法への準拠を実証するための自主的なツールとして機能する。
What are the next steps after the Code is published? 規範の公表後の次のステップは?
After the publication, Member States and the Commission will assess the adequacy of the Code. 公表後、加盟国と欧州委員会は、規範の妥当性を評価する。
The Commission will complement the Code with the guidelines on general-purpose AI models, which will be published ahead of the entry into application of the rules for providers of general-purpose AI models. The guidelines will clarify, for example, who is in and out of scope of the AI Act’s obligations for providers of general-purpose AI models. 欧州委員会は、汎用 AI モデルに関するガイドラインを実践規範に補足する。このガイドラインは、汎用 AI モデルのプロバイダに対する規則の適用開始に先立ち公表される。このガイドラインでは、例えば、汎用 AI モデルのプロバイダに対する AI 法の義務の対象となる者および対象とならない者を明確にする。
From 2 August 2025 onwards, providers placing general-purpose AI models on the market must comply with their respective AI Act obligations. Providers must notify the AI Office of general-purpose AI models with systemic risk to be placed on the EU market without delay. In the first year from 2 August 2025 onwards, the AI Office will offer to collaborate closely in particular with providers who adhere to the Code to ensure that models can continue to be placed on the EU market without delays. In particular, if these providers do not fully implement all commitments immediately after signing the Code, the AI Office will not consider them to have broken their commitments under the Code and will not reproach them for violating the AI Act. Instead, in such cases, the AI Office will consider them to act in good faith and will be ready to collaborate to find ways to ensure full compliance. However, from 2 August 2026 onwards, the Commission will enforce full compliance with all obligations for providers of general-purpose AI models with fines. 2025 年 8 月 2 日以降、汎用 AI モデルを上市するプロバイダは、それぞれの AI 法の義務を遵守しなければならない。プロバイダは、EU 市場に上市される、システミックリスクのある汎用 AI モデルについて、AI 事務局に遅滞なく通知しなければならない。2025 年 8 月 2 日から 1 年間は、AI 事務局は、実践規範を遵守するプロバイダと特に緊密な連携を取り、モデルが EU 市場に遅滞なく上市され続けるよう努める。特に、これらのプロバイダが実践規範の署名直後にすべての約束を完全に実施しなかった場合、AI 事務局は、そのプロバイダが実践規範に基づく約束に違反したとはみなさず、AI 法違反として非難することはない。その代わりに、AI 事務局は、そのプロバイダが誠意をもって行動しているとみなして、完全な遵守を確保するための方法を見つけるために協力する用意がある。ただし、2026 年 8 月 2 日以降、欧州委員会は、汎用 AI モデルのプロバイダに対して、すべての義務の完全な遵守を罰金によって強制する。
Models placed on the market before 2 August 2025 must comply with the AI Act obligations by 2 August 2027. 2025 年 8 月 2 日より前に上市されたモデルは、2027 年 8 月 2 日までに AI 法の義務を遵守しなければならない。
How will the Code be updated? 実践規範はどのように更新されるのか?
The Chairs and Vice-Chairs have written the Code to be as future-proof as possible. However, the rapid evolution of AI technology poses a challenge. Even with the future-proof design, the Code will still require periodic updates. 議長および副議長は、実践規範を可能な限り将来も通用するものにするよう作成した。しかし、AI 技術の急速な進化は課題となっている。将来を見据えた設計であっても、実践規範は定期的な更新が必要になるだろう。
The AI Office may facilitate formal updates to the Code in response to technological developments, changes in the risk landscape, or experience with the application of the AI Act rules. The AI Office will review the Code at least every two years, and it may propose a streamlined process for reviews and updates as needed. AI 事務局は、技術の発展、リスクの状況の変化、AI 法規則の適用経験に応じて、実践規範の正式な更新を促進することができる。AI 事務局は、少なくとも 2 年ごとに実践規範を見直し、必要に応じて、見直しと更新のための合理化されたプロセスを提案することができる。
What role will the Commission guidelines on general-purpose AI models play? 汎用 AI モデルに関する欧州委員会のガイドラインはどのような役割を果たすのか?
The Commission guidelines on general-purpose AI models will provide clarity on the scope of obligations for providers. 汎用 AI モデルに関する欧州委員会のガイドラインは、プロバイダの義務の範囲を明確にするものだ。
A public consultation in April and May collected stakeholder input on key concepts. Using feedback from providers, experts, and other stakeholders, the AI Office is finalising the guidelines, which will be published before the general-purpose AI obligations’ entry into application. The guidelines will clarify three key points: 4 月と 5 月に実施された公開協議では、主要な概念についてステークホルダーからの意見が収集された。AI 事務局は、プロバイダ、専門家、その他のステークホルダーからのフィードバックを参考にし、ガイドラインの最終案をまとめている。このガイドラインは、汎用 AI 義務の適用開始前に公表される予定である。このガイドラインでは、3 つの重要な点が明確になる。
・When a model is a general-purpose AI model ・モデルが汎用 AI モデルである場合
・When a model is a general-purpose AI model with systemic risk ・モデルが、システミックリスクを伴う汎用 AI モデルである場合
・Who is considered the provider of such a model ・そのようなモデルの提供者とみなされる者
Where to find further information? 詳細情報
Press release プレスリリース
The Code of Practice on GPAI GPAI に関する実践規範
More information on the Code of Practice on GPAI GPAI に関する実践規範の詳細
More information on the Chairs and Vice-Chairs 議長および副議長に関する詳細

 

 

 

 


 

⚫︎ まるちゃんの情報セキュリティ気まぐれ日記

・2025.04.11 欧州委員会 AI大陸行動計画 (2025.04.09)

・2025.03.13 欧州委員会 AI法における汎用AIモデル - Q&A (2025.03.10)

・2025.02.08 欧州委員会 規則(EU)2024/1689(AI法)が定める人工知能の禁止行為に関する欧州委員会ガイドライン

・2024.12.22 欧州委員会 汎用AI実践規範の第2ドラフトを発表 (2024.12.19)

・2024.12.06 欧州委員会 AI法における汎用AIモデル - Q&A (2024.11.20)

・2024.11.17 欧州委員会 汎用AI実践規範の最初のドラフトを発表 (2024.11.14)

・2024.10.30 欧州 AI法の調和標準の策定について...

・2024.09.08 欧州評議会 AI条約の署名開始... (2024.09.05)

・2024.08.05 欧州AI法が施行された... (2024.08.01)

・2024.07.19 ドイツ BfDI フランス CNIL オランダ AP EU AI法関連

・2024.07.16 EU 2024.07.12にAI法がEU官報に掲載された

・2024.05.22 EU 欧州理事会がAI法を承認...まもなく発効されますね...

・2024.03.19 欧州議会 AI法 (2024.03.13) とサイバーレジリエンス法を採択 (2024.03.12)

・2023.12.10 欧州委員会、欧州議会がAI法について政治的合意

・2022.12.08 EU理事会 AI法に関する見解を採択

・2022.09.30 欧州委員会 AI責任指令案

・2021.12.05 欧州理事会 AI法改正案を欧州議会に提出

・2021.08.08 EU議会 BRIEFING 人工知能法 at 2021.07.26

・2021.04.24 欧州委員会がAIへの規制を提案 → 欧州データ保護官は歓迎するけど、公共空間での遠隔生体認証についての規制も入れてね

 

 

 

|

« 米国 NIST SP 1334( 初期公開ドラフト) OT 環境における可搬保管媒体のサイバーセキュリティリスクの軽減 (2025.07.15) | Main | 米国 ホワイトハウス AI行動計画(2025.07.23) »

Comments

Post a comment



(Not displayed with comment.)


Comments are moderated, and will not appear on this weblog until the author has approved them.



« 米国 NIST SP 1334( 初期公開ドラフト) OT 環境における可搬保管媒体のサイバーセキュリティリスクの軽減 (2025.07.15) | Main | 米国 ホワイトハウス AI行動計画(2025.07.23) »