AIと統合された職場における労働者の福祉を促進するためのフレームワーク

こんにちは、丸山満彦です。

Partnership on AIから「AIと統合された職場における労働者の福祉を促進するためのフレームワーク」に関する報告書が公表されていますね。。。

● Partnership on AI

・2020.09.15 Framework for Promoting Workforce Well-being in the AI-Integrated Workplace

・[PDF] Full Report

---

Table Of Contents

Executive Summary

Acknowledgements

About The Partnership on AI

Introduction

• Applications of AI in the Workplace
• Ethical Questions Associated with AI Integration
• Why Workforce Well-being
• About the Framework

PART 1

Six Pillars of Workforce Well-being and AI Impact

1. Human Rights
2. Physical Well-being
3. Financial Well-being
4. Intellectual Well-being
5. Emotional Well-being
6. Purpose and Meaning

Summary Chart: AI Impact and the Six Pillars of Workforce Well-being

PART 2

Framework for Promoting Workforce Well-being in the AI Integrated Workplace

1. Key Considerations
2. Recommendations
3. Discussion Questions
4. Worker Survey

Conclusion and Future Work

Appendix 1 - Definitions

Appendix 2 - Literature Review

---

Continue reading "AIと統合された職場における労働者の福祉を促進するためのフレームワーク"

英国 データ保護委員会 アカウンタビリティ フレームワーク

こんにちは、丸山満彦です。

英国のデータ保護委員会（Information Commissioner's Office : ICO）がアカウントビリティー フレームワークを公表していますね。大作です！

● UK-ICO

・2020.09.10 Blog: Accountability Framework: demonstrating your compliance

Ian Hulme, Director of Regulatory Assurance discusses the launch of our new Accountability Framework and how organisations can take part in the next stage of its development. 

・Accountability Framework

• Introduction to the Accountability Framework
  
  
• 01. Leadership and oversight
  • Organisational structure
  • Whether to appoint a DPO
  • Appropriate reporting
  • Operational roles
  • Oversight groups
  • Operational group meetings
    
    
• 02. Policies and procedures
  • Direction and support
  • Review and approval
  • Staff awareness
  • Data protection by design and by default
    
    
• 03. Training and awareness
  • All-staff training programme
  • Induction and refresher training
  • Specialised roles
  • Monitoring
  • Awareness raising
    
    
• 04. Individuals’ rights
  • Informing individuals and identifying requests
  • Resources
  • Logging and tracking requests
  • Timely responses
  • Monitoring and evaluating performance
  • Inaccurate or incomplete information
  • Erasure
  • Restriction
  • Data portability
  • Rights related to automated decision-making and profiling
  • Individual complaints
    
    
• 05. Transparency
  • Privacy notice content
  • Timely privacy information
  • Effective privacy information
  • Automated decision-making and profiling
  • Staff awareness
  • Privacy information review
  • Tools supporting transparency and control
    
    
• 06. Records of processing and lawful basis
  • Data mapping
  • Record of processing activities (ROPA)
  • ROPA requirements
  • Good practice for ROPAs
  • Documenting your lawful basis
  • Lawful basis transparency
  • Consent requirements
  • Reviewing consent
  • Risk-based age checks and parental or guardian consent
  • Legitimate interest assessment (LIA)
    
    
• 07. Contracts and data sharing
  • Data sharing policies and procedures
  • Data sharing agreements
  • Restricted transfers
  • Processors
  • Controller-processor contract requirements
  • Processor due diligence checks
  • Processor compliance reviews
  • Third-party products and services
  • Purpose limitation
    
    
• 08. Risks and data protection impact assessments (DPIAs)
  • Identifying, recording and managing risks
  • Data protection by design and by default
  • DPIA policy and procedures
  • DPIA content
  • DPIA risk mitigation and review
    
    
• 09. Records management and security
  • Creating, locating and retrieving records
  • Security for transfers
  • Data quality
  • Retention schedule
  • Destruction
  • Information asset register
  • Rules for acceptable software use
  • Access control
  • Unauthorised access
  • Mobile devices, home or remote working and removable media
  • Secure areas
  • Business continuity, disaster recovery and back-ups
    
    
• 10. Breach response and monitoring
  • Detecting, managing and recording incidents and breaches
  • Assessing and reporting breaches
  • Notifying individuals
  • Reviewing and monitoring
  • External audit or compliance check
  • Internal audit programme
  • Performance and compliance information
  • Use of management information

---

 

2024年までに戦闘機でAIパイロットをテストすることを含め、米国はAIの軍事利用で世界をリードする by エスパー国防長官

こんにちは、丸山満彦です。

米国のエスパー国防長官が、9月9日と10日に開催されたDoDのAIシンポジウム（2020 Department of Defense Artificial Intelligence Symposium and Exposition）でエスパー国防長官が、「2024年までに戦闘機でAIパイロットをテストすることを含め、米国は人工知能の軍事利用で世界をリードする」と述べたようですね。

● DoDのAIブログ

・2020.09.10 Making AI Real: Department of Defense AI Symposium Day 1

では、そこまで具体的に書かれていませんが、

● Breaking Defense

・2020.09.09 AI To Fly In Dogfight Tests By 2024 by SYDNEY J. FREEDBERG JR.

After an AI beat humans 5-0 in AlphaDogfight simulations this summer, Mark Esper announced, a future version will be installed in actual airplanes for “a real-world competition.” But military AI will adhere to strict ethical limits, he said.

によると、

The US, he vowed, will lead the world on the military use of artificial intelligence – including testing an AI pilot in a fighter by 2024.

ということのようです。。。

■ 参考
● 2020 Department of Defense Artificial Intelligence Symposium and Exposition

● Breaking Defense
・2020.08.20 AI Slays Top F-16 Pilot In DARPA Dogfight Simulation by THERESA HITCHEN

"It's a giant leap," said DARPA's Justin (call sign "Glock") Mock.

● まるちゃんの情報セキュリティ気まぐれ日記

・2020.08.24 AIがDARPAドッグファイトシミュレーションでトップレベルのF-16パイロットを倒す?

 

マイクロソフトがDeepfake検出ツールを発表してました。。。

こんにちは、丸山満彦です。

マイクロソフトのブログで、Deepfakeを検出するツールについての記事がありました。。。

● Microsoft - Microsoft On the Issues

・2020.09.01 New Steps to Combat Disinformation

大統領選が近づいていますからね。。。

Fake動画を作る側とそれを見破る側の「イタチごっこ」が始まりますね。。。

ちなみに、英語では”Cat and mouse”　(^^)

Source Microsoft

また、Microsoftは、Deepfakeに関するキャンペーンもしているようですね。

●Stop Deepfake

---

 ■ 報道等

● esst - WeLiceSecurity
・2020.09.03 Microsoft debuts deepfake detection tool by Amer Owaida

As the US presidential election nears, the company’s new tech should also help assure people that an image or video is authentic

● TechCranch
・2020.09.02 Microsoft launches a deepfake detector tool ahead of US election by Natasha Lomas

● BBC
・2020.09.02 Deepfake detection tool unveiled by Microsoft by Leo Kelion

Microsoft has developed a tool to spot deepfakes - computer-manipulated images in which one person's likeness has been used to replace that of another.

脳とコンピュータの結合（Brain Computer Interface）

こんにちは、丸山満彦です。

イーロン・マスク氏が2016年に創業した米Neuralink社が昨年披露したチップより大幅に改良された「LINK VO.9」（以下「LINK」）、LINKを脳に移植するための外科用ロボット、既にLINKを埋め込んだブタによるデモを披露し、話題になっていますね。。。

● IT Media
・2020.08.30 イーロン・マスクのNeuralink、脳埋め込みチップのブタでのデモで進捗報告

この記事によると、LINKは、7月に米食品医薬品局（FDA）のBreakthrough Devices Programの認定を受けていて、人間への移植準備はほぼできており、安全性のテストを重ねて許可を待っている状態のようですね。

さて、このような技術は、Brain Machine Interface (BMI)とか、Brain Computer Interface (BCI)とか言われていますが、こういう研究を一番進めているのは、米軍なんでしょうかね。。。

さて、米国ではこの発表の1日前の2020.08.27にRand研究所[wikipedia]からBCIに関する報告書が公表されています。どのような場面で活用ができそうかという話は当然として、潜在的なリスク（Potencial Risks）にも相当触れられていますね。。。

● Rand Corporation

・2020.08.27 (article) Brain-Computer Interfaces Are Coming. Will We Be Ready?

・2020.08.27 (report) Brain-Computer Interfaces - U.S. Military Applications and Implications, An Initial Assessment - by Anika Binnendijk, Timothy Marler, Elizabeth M. Bartels

・本文[Online] [PDF] [Downloaded]

目次は次の通り

---

Summary

Introduction
 Human-Machine Teaming
 Approach

Technology Summary
 Introduction
 Review
 Development Directions and Technical Challenges
  Development Directions
  Technical Challenges and Risks

OPERATIONAL CONSIDERATIONS
 Introduction
 BCI and the Future Warfighter
 Potential BCI Applications in Future Combat
 Existing Concepts for BCI Combat Applications

TESTING BCI CAPABILITIES THROUGH NATIONAL SECURITY GAMING
 A Projected BCI Toolbox
 Testing the Operational Relevance of BCI Capabilities
 Insights on the Use of BCI from Game Play
  Vignette 1: Clearing a Building
  Vignette 2: Ambush and Casualty Evacuation
  Additional Use Cases
  Summary of Game Findings

POTENTIAL RISKS
 Operational Vulnerabilities  
  New Potential Points of Failure
  Adversary Access to New Information
  New Areas of Exposure to Harm or Influence
 Institutional Vulnerabilities
  Trust
  Erosion of Unit Cohesion
  Erosion of Unit Leadership
 Ethical and Legal Risks
  Responsibility to the BCI Operator
  Responsibility of the BCI Operator

CONCLUSIONS AND RECOMMENDATIONS
 Summary and Primary Findings
 Recommendations
  Expand Analyses to Illuminate Operational Relevance and Vulnerabilities
  Address the Trust Deficit
  Collaborate and Anticipate
  Plan Ahead for BCI Institutional Implications

APPENDIX. GAME DESIGN AND EXECUTION
 Vignette Selection
 Game Process

NOTES
REFERENCES
ACKNOWLEDGMENTS

---

報道等

● Braking Defense
・2020.08.27 DoD Needs New Policies, Ethics For Brain-Computer Links (Jacked-In Troops?) By   THERESA HITCHENS

"BCI is not just science fiction; it has viable practical applications, but there is much more work that needs to be completed before it becomes mainstream and commercial," RAND's Tim Marler says.

Continue reading "脳とコンピュータの結合（Brain Computer Interface）"

人工知能と感情知性に関する倫理（2020.07.30）

こんにちは、丸山満彦です。

「人工知能と感情知性に関する倫理」についてのペーパーがPartnership on AI [wikipedia]から出されていました。。。7月30日だから1ヶ月ほど前です。。。

関連する多くの論文等が引かれていて参考になりそうです。感情はプライバシーに深く関連してくるので、しっかりとした議論をしてコンセンサスをとっていく必要がありそうですね。。。

● Partnership on AI

・2020.07.30 (Paper) The Ethics of AI and Emotional Intelligence

・[pdf] Full paper [Downloaded]

-----

Contents

Executive Summary

Introduction

Affective Computing Definition and Language Problems

• Affective computing and AI
• Language problems in public discussions

Sensors, Inputs, Inferences and Applications

• Breadth of signals in use
• Examples of affective computing input data sources

Inferences and Applications

• Detecting, treating or assisting with disease or disability
• Agriculture
• Social robots
• Education and audience engagement
• Gaming, movies and entertainment
• Advertising and retail
• Hiring and employment
• Chatbots, call centers, and home/auto voice assistants
• Wearables and stress relief
• Automotive and industrial safety
• Threat detection/intervention and law enforcement
• Communities, politics, and social networks

Questions for Discussion

• Thinking big
• How does affective computing fit in existing frameworks?
• Human vs. machine
• Accuracy and inclusivity
• Privacy and other rights
• Autonomy and best interest
• Transparency and communications

Question Exploration as a Tool for Evaluating Ethics Risk

• Who has access to the inferences in depression detection and why does it matter?
• What are the biggest opportunities and benefits of affective computing for society?
• What are the greatest risks of affective computing for society?

COVID-19, Black Lives Matter Protests, and Affective Computing

• AI could alleviate COVID-19 mental health problems
• Social robots could help reduce human proximity and disease transmission
• The shift to remote work and education raises issues of privacy and control and may increase demand for affective computing technology
• Pandemic health monitoring and affective computing share the privacy risks of tracking biometric, health, and location data
• The Black Lives Matter protests and affective computing raise overlapping bias issues
• The Black Lives Matter protests are changing the law around law enforcement camera use and facial analysis

Conclusion

Acknowledgments

---

Continue reading "人工知能と感情知性に関する倫理（2020.07.30）"

AIがDARPAドッグファイトシミュレーションでトップレベルのF-16パイロットを倒す?

こんにちは、丸山満彦です。

AIがDARPAドッグファイトシミュレーションでトップレベルのF-16パイロットを倒したようですね。。。

戦闘機に要求される条件も変わってくるのかもしれませんね。。。

● YouTube - DARPA chanel

・2020.08.22 AlphaDogfight Trials Final Event

人間とAIの戦いは4時間40分ごろくらいからですね。。。。

● Braking Defense

・2020.08.20 AI Slays Top F-16 Pilot In DARPA Dogfight Simulation

"It's a giant leap," said DARPA's Justin (call sign "Glock") Mock.

● このブログ

・2020.08.10 AIと将来の軍事力についての中国の視点

・2020.06.16「米海兵隊はAIを理論から実践に移す」というブログ

・2020.06.09 米空軍は、2021年7月にAIが操縦する戦闘機と人間が操縦する戦闘機でドッグファイトをすることを考えているようですね。。。

・2020.05.04 米国国防省と人工知能（戦略と倫理）

・2020.02.13 無人兵器・・・人工知能はついているかも・・・

NISTはAIに自己説明を求める？（説明可能な人工知能の4原則）

こんにちは、丸山満彦です。

NISTが説明可能な人工知能の4原則のドラフトを公表し意見を募集していますね。。。

● NIST

・2020.08.18 (News) NIST Asks A.I. to Explain Itself

Technical agency proposes four fundamental principles for judging how well AI decisions can be explained.

・[PDF] Draft NISTIR 8312 Four Principles of Explainable Artificial Intelligence 

---

Abstract

We introduce four principles for explainable artificial intelligence (AI) that comprise the fundamental properties for explainable AI systems. They were developed to encompass he multidisciplinary nature of explainable AI, including the fields of computer science, engineering, and psychology. Because one size fits all explanations do not exist, different users will require different types of explanations. We present five categories of explanation and summarize theories of explainable AI. We give an overview of the algorithms in the field that cover the major classes of explainable algorithms. As a baseline comparison, we assess how well explanations provided by people follow our four principles. This assessment provides insights to the challenges of designing explainable AI systems.

---

４つの原則とは

1. Explanation: 
   AI systems should deliver accompanying evidence or reasons for all outputs.
   
   
2. Meaningful: 
   Systems should provide explanations that are understandable to individual users.
   
   
3. Explanation Accuracy: 
   The explanation should correctly reflect the system’s process for generating the output.
   
   
4. Knowledge Limits: 
   The system only operates under conditions for which it was designed or when the system reaches a sufficient level of confidence in its output.

ざっというと

1. 説明：
   AIシステムは、全ての出力に付随する証拠や理由を提供する必要がある。
2. 有意味：
   システムは、個々の利用者が利用できる説明を提供する必要がある。
   
   
3. 説明の正確性：
   説明は、出力を生成するためのシステムプロセスを正しく反映している。
   
   
4. 知識の限界：
   システムは、それが設計された条件下でのみ又は、システムがその出力が十分に信頼できる状況になった場合にのみ動作する。
   
   

という感じでしょうかね。。。

目次です。

---

Table of Contents

1 Introduction

2 Four Principles of Explainable AI
 2.1 Explanation
 2.2 Meaningful
 2.3 Explanation Accuracy
 2.4 Knowledge Limits

3 Types of Explanations

4 Overview of principles in the literature

5 Overview of Explainable AI Algorithms
 5.1 Self-Explainable Models
 5.2 Global Explainable AI Algorithms
 5.3 Per-Decision Explainable AI Algorithms
 5.4 Adversarial Attacks on Explainability

6 Humans as a Comparison Group for Explainable AI
 6.1 Explanation
 6.2 Meaningful
 6.3 Explanation Accuracy
 6.4 Knowledge Limits

7 Discussion and Conclusions

References

---

Continue reading "NISTはAIに自己説明を求める？（説明可能な人工知能の4原則）"

NPO デジタル・フォレンジック研究会のコラム by 丸山満彦

こんにちは、丸山満彦です。

デジタル・フォレンジック研究会にコラムを載せました。

● IDF - Column

・2020.08.17 第６２７号コラム：「若者のサイバー犯罪を無くしたい。。。」

若者がサイバー犯罪の被害者にならないようにすることはもちろん重要ですが、逆に加害者にならないようにすることも重要という思いです。。。

ーーーーー

私が書いた「NPO デジタル・フォレンジック研究会」の「コラム」の一覧

＃	No	Date	Title
23	627	2020.08.17	若者のサイバー犯罪を無くしたい。。。
22	600	2020.02.03	デジタルフォレンジックスと多様性
21	578	2019.08.26	未来を考えようと思うとき、人は過去を振り返る
20	551	2019.02.11	とらわれずに物事をみつめる
19	521	2018.07.09	ＡＩは科学捜査を騙せるか？
18	493	2017.12.18	セキュリティ・デバイド？
17	474	2017.08.07	『デジタル・フォレンジック』という言葉を今更考える
16	451	2017.02.20	相手を知ることが重要
15	425	2016.08.15	本質を理解する
14	383	2015.10.12	名ばかりCSIRTで良いのか？
13	357	2015.04.13	ＩｏＴ時代は明るいか暗いか
12	335	2014.11.03	頭を下げるのは社長です
11	308	2014.04.30	標的型攻撃には内部不正対応が重要？
10	286	2013.11.14	セキュリティガバナンスはできる範囲だけやればよいのか？
09	261	2013.05.23	セキュリティの基本はずっとかわっていない
08	240	2012.12.25	さらに組織化が進むサイバー攻撃集団
07	207	2012.05.10	外部から侵入されている想定で情報セキュリティを考える
06	173	2011.09.08	想定外に対応するのが危機管理ではないか
05	139	2011.01.13	データ分析を使った不正発見手法
04	131	2010.11.11	発見的統制の重要性
03	084	2009.12.10	クラウドコンピューティングがもたらす光と影
02	058	2009.06.11	不正をさせない
01	021	2008.09.25	ニーズとシーズ、目的と手段

DeepfakeについてのNATO, CSET, Partnership on AI, GAOの報告書（少し前ですが・・・）

こんにちは、丸山満彦です。

Deepfakeについての公表物がいくつかあるので、備忘録・・・

Deepfakeを作る技術と見破る技術のイタチごっこ, cat-and-mouse gameが続くのでしょうね。。。

映画とかでは面白い技術とは思いますが、SNSで作為を持って流されるなども考えられるので、何らかの規制が将来は必要となるかもしれませんね。。。

いくつか並べてみましたが、NATOとCSETの著者は同じでTim Hwang氏ですね。。。

● NATO Strategic Communications Center of Excellence Riga、Latvia

・2020.06.05 New study offers insight into future of deepfake technology

・[PDF] Deepfakes - Primer and Forecast by Tim Hwang

● CENTER for SECURITY and EMERGING TECHNOLOGY (CSET)

・2020.07 (Analysis) Deepfakes: A Grounded Threat Assessment by Tim Hwang

・[PDF] Deepfakes: A Grounded Threat Assessment

The rise of deepfakes could enhance the effectiveness of disinformation efforts by states, political parties and adversarial actors. How rapidly is this technology advancing, and who in reality might adopt it for malicious ends? This report offers a comprehensive deepfake threat assessment grounded in the latest machine learning research on generative models.

● Partnership on AI

・2020.03.12 (NEWS) A Report on the Deepfake Detection Challenge by CLAIRE LEIBOWICZ

・[PDF] The Deepfake Detection Challenge: Insights and Recommendations for AI and Media Integrity

● GAO

・2020.02.20 SCIENCE & TECH SPOTLIGHT: Deepfakes GAO-20-379SP

・[PDF] Full report

その他

● Reserchgate.net

・2019.09 Deep Learning for Deepfakes Creation and Detection: A Survey by Thanh Thi Nguyen and Cuong M. Nguyen

Abstract

Deep learning has been successfully applied to solve various complex problems ranging from big data analytics to computer vision and human-level control. Deep learning advances however have also been employed to create software that can cause threats to privacy, democracy and national security. One of those deep learning-powered applications recently emerged is "deepfake". Deepfake algorithms can create fake images and videos that humans cannot distinguish them from authentic ones. The proposal of technologies that can automatically detect and assess the integrity of digital visual media is therefore indispensable. This paper presents a survey of algorithms used to create deepfakes and, more importantly, methods proposed to detect deepfakes in the literature to date. We present extensive discussions on challenges, research trends and directions related to deepfake technologies. By reviewing the background of deepfakes and state-of-the-art deepfake detection methods, this study provides a comprehensive overview of deepfake techniques and facilitates the development of new and more robust methods to deal with the increasingly challenging deepfakes.

● Springer Link

・2020.08.06 The Epistemic Threat of Deepfakes by Don Fallis

 ・[PDF]

Abstract

Deepfakes are realistic videos created using new machine learning techniques rather than traditional photographic means. They tend to depict people saying and doing things that they did not actually say or do. In the news media and the blogosphere, the worry has been raised that, as a result of deepfakes, we are heading toward an “infopocalypse” where we cannot tell what is real from what is not. Several philosophers (e.g., Deborah Johnson, Luciano Floridi, Regina Rini) have now issued similar warnings. In this paper, I offer an analysis of why deepfakes are such a serious threat to knowledge. Utilizing the account of information carrying recently developed by Brian Skyrms (2010), I argue that deepfakes reduce the amount of information that videos carry to viewers. I conclude by drawing some implications of this analysis for addressing the epistemic threat of deepfakes.

・2020.06.23 DeepFake Video Detection: A Time-Distributed Approach by Amritpal Singh, Amanpreet Singh Saimbhi, Navjot Singh & Mamta Mittal

・[PDF]

Abstract

Recent developments in machine learning algorithms have led to the generation of forged videos having remarkable quality, which are indistinguishable from real videos. This can fatally affect the way in which one perceives the information available digitally. Thus, this paper aims to efficiently and holistically detect manipulated videos generated using DeepFake, which is the most effective deep learning powered technique developed so far by the researchers. Arduous efforts have been put to detect the forgery in still images, but the authors leveraged the spatio-temporal features of the videos by taking sequences of frames as input to the model. Furthermore, the authors have proposed an architecture which took advantage of lower-level features around regions of interest as well as discrepancies across multiple frames. Experiments have been performed on the Deep Fake Detection Challenge dataset of ≈≈ 470 GB in size, and it has been observed that the proposed approach yielded a test accuracy score of 97.6%.

---