~7 min read · 2026-05-21

DAI Pillar A — Governance 调研

注：此次调研只跑到证据采集阶段，未生成正式报告正文。以下为采集到的来源材料。

Batch 1 sources - regulatory + stats verification (all 2026-05-21)

EU AI Act#

• VERIFIED: EU AI Act enters full applicability 2026-08-02
• GPAI Code of Practice published 2025-07-10
• Penalties: greater of €15M or 3% global turnover from 2026-08-02
• URL: https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
• URL: https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai

Gartner 80%#

• VERIFIED: Gartner press release 2024-11-13
• "80% of Non-Executive Directors Believe Current Board Practices and Structures Are Inadequate to Oversee AI"
• URL: https://www.gartner.com/en/newsroom/press-releases/2024-11-13-gartner-says-80-percent-of-non-executive-directors-believe-current-board-practices-and-structures-are-inadequate-to-oversee-ai

McKinsey 39%#

• VERIFIED: McKinsey "Elevating board governance through AI posture and archetypes"
• "As of 2024, only 39% of Fortune 100 companies disclosed any form of board oversight of AI"
• 2025 update: 40% disclosed board-level committee for AI oversight (up from 11% in 2024)
• URL: https://www.mckinsey.com/capabilities/mckinsey-technology/our-insights/the-ai-reckoning-how-boards-can-evolve

Bain "What Boards Need to Do About AI"#

• VERIFIED: Published 2025-03-07
• URL: https://www.bain.com/insights/what-boards-need-to-do-about-ai/

IAPP Agentic AI 93%/7%#

• PARTIAL: Could not verify exact 93%/7% number in this batch
• Related IAPP/Transcend reports show ~75% giving agentic AI access to data, only 20% have tested response plan
• URL: https://iapp.org/resources/article/ai-governance-in-the-agentic-era
• Need follow-up search for exact stat

OWASP Top 10 Agentic#

• VERIFIED: Released 2025-12-09 (as user said 2025-12)
• OWASP GenAI Security Project, 100+ contributors, framework version "for 2026"
• URL: https://genai.owasp.org/resource/owasp-top-10-for-agentic-applications-for-2026/
• 10 risk categories incl. Agent Goal Hijack, Tool Misuse, Identity/Privilege Abuse, Supply Chain, Code Execution, Context Manipulation, Inter-agent, Over-reliance, Rogue Agents
• Key principle: "least agency"

NIST AI Agents RFI#

• VERIFIED: Published in Federal Register 2026-01-08 (docket NIST-2025-0035)
• From CAISI (Center for AI Standards and Innovation)
• Comment period closed 2026-03-09
• URL: https://www.federalregister.gov/documents/2026/01/08/2026-00206/request-for-information-regarding-security-considerations-for-artificial-intelligence-agents
• URL: https://www.nist.gov/caisi/ai-agent-standards-initiative

Human-in-the-loop paradox PMC 2024#

• VERIFIED: PLoS One 2024 study, PMC10857587
• Title: "Putting a human in the loop: Increasing uptake, but decreasing accuracy of automated decision-making"
• URL: https://pmc.ncbi.nlm.nih.gov/articles/PMC10857587/

Batch 2-4 faculty verification (all checked 2026-05-21)

Academics verified#

1. 1. Cary Coglianese — Penn Carey Law, Edward B. Shils Prof of Law, Director Penn Program on Regulation. Recent: "Leashes, Not Guardrails: A Management-Based Approach to AI Risk Regulation" (Risk Analysis 2024-25); "Strategies for Regulating Artificial Intelligence" (2025); PA AI Advisory Committee 2026. URL: https://www.law.upenn.edu/faculty/ccoglian/

1. 2. Anu Bradford — Columbia Law, Henry L. Moses Prof. Books: "Digital Empires" (Oxford 2023), "Brussels Effect" (2020). URL: https://www.law.columbia.edu/faculty/anu-bradford

1. 3. Margot Kaminski — CU Boulder Law, Associate Prof, Director Privacy Initiative at Silicon Flatirons. Appointed to CO AI Impact Task Force 2024. Recent: "Regulating the Risks of AI" Boston U Law Rev 2023; "Binary Governance" (GDPR). URL: https://www.colorado.edu/law/people/margot-kaminski

1. 4. Jonas Tallberg — Stockholm U, Prof of Political Science, Co-Director Stockholm Center on Global Governance. Recent: "Global Governance of AI: Next Steps for Empirical and Normative Research" (Int'l Studies Review 2023). URL: https://www.su.se/english/profiles/jtall-1.183532

1. 5. Lawrence Lessig — Harvard Law, Roy L. Furman Prof of Law. Cited in 138 Harv L Rev 1562 "Resetting Antidiscrimination Law in the Age of AI" (2025); Berkman advisory board for AI & Law Initiative. URL: https://hls.harvard.edu/faculty/lawrence-lessig/

1. 6. Helen Nissenbaum — Cornell Tech, Andrew H. & Ann R. Tisch Prof of Info Science, Founding Director Digital Life Initiative. URL: https://nissenbaum.tech.cornell.edu/

1. 7. Ryan Calo — UW Law, Lane Powell Prof; co-founder UW Tech Policy Lab. "AI Policy: Primer and Roadmap". Testified Senate 4x. URL: https://www.law.uw.edu/directory/faculty/calo-ryan

1. 8. Sandra Wachter — Oxford Internet Institute, Prof of Tech and Regulation. Counterfactual Explanations adopted by Google/IBM/Vodafone. URL: https://www.oii.ox.ac.uk/people/profiles/sandra-wachter/

1. 9. Daniel Ho — Stanford Law, William Benjamin Scott & Luna M. Scott Prof, Director RegLab. NAIAC member, advisor DOL. URL: https://law.stanford.edu/daniel-e-ho/

1. 10. Gillian Hadfield — Johns Hopkins, Bloomberg Distinguished Prof of AI Alignment and Governance (named June 2025). Former OpenAI Senior Policy Advisor 2018-2023. Concept: "regulatory markets". URL: https://bdp.jhu.edu/bd-professors/gillian-hadfield/

1. 11. Angela Huyue Zhang — USC Gould School of Law, Prof of Law. Book: "High Wire: How China Regulates Big Tech" (Oxford 2024). URL: https://gould.usc.edu/faculty/profile/zhang-angela/

1. 12. John Villasenor — UCLA, Prof of Engineering / Law / Public Policy / Management; Non-resident Senior Fellow Brookings. URL: https://law.ucla.edu/faculty/faculty-profiles/john-villasenor

Think tank / civil society verified#

1. 13. Helen Toner — CSET Georgetown, Interim Executive Director (since Sep 2025). Time 100 AI 2024. Testified Senate Judiciary April 2026. URL: https://cset.georgetown.edu/staff/helen-toner/

1. 14. Allan Dafoe — Google DeepMind Director of Frontier Safety and Governance; Senior Advisor Oxford Martin AI Governance Initiative. Founder GovAI. URL: https://www.governance.ai/team/allandafoe

1. 15. Darrell M. West — Brookings Senior Fellow, Center for Tech Innovation. Co-editor TechTank. Book: "Turning Point: Policymaking in the Era of AI" (Brookings Press 2020). URL: https://www.brookings.edu/people/darrell-m-west/

1. 16. Russell Wald — Stanford HAI Executive Director (formerly Director of Policy). URL: https://hai.stanford.edu/people/russell-wald

1. 17. Marc Rotenberg — Founder & Executive Director CAIDP; teaches at Georgetown Law; former founder/director EPIC. URL: https://www.caidp.org/about-2/team/

1. 18. Janet Haven — Executive Director Data & Society; NAIAC member 2022-2025 (chaired Rights/Trust/Safety WG); UN AI advisory body. URL: https://datasociety.net/people/haven-janet/

1. 19. Marietje Schaake — Stanford Cyber Policy Center International Policy Director; Stanford HAI International Policy Fellow. Former MEP 2009-2019. Book: "The Tech Coup" (2024). UN High-Level Advisory Body on AI. URL: https://cyber.fsi.stanford.edu/people/marietje-schaake

1. 20. Matt Sheehan — Carnegie Endowment Senior Fellow Asia Program. Paper series: "China's AI Regulations and How They Get Made". URL: https://carnegieendowment.org/people/matt-sheehan

1. 21. Alondra Nelson — Institute for Advanced Study, Harold F. Linder Chair. Former OSTP Acting Director 2021-2023. Led Blueprint for AI Bill of Rights. URL: https://www.ias.edu/sss/faculty/nelson

1. 22. Suresh Venkatasubramanian — Brown U, Prof of CS & Data Science, Director CNTR. Co-author of AI Bill of Rights. URL: https://vivo.brown.edu/display/suresh

1. 23. Beth Simone Noveck — Northeastern (not NYU now), Director Burnes Center for Social Change; former NJ Chief AI Strategist; former Deputy CTO White House. URL: https://burnes.northeastern.edu/people/beth-noveck/

1. 24. Inioluwa Deborah Raji — UC Berkeley researcher; Academic Fellow Leadership Conference on Civil and Human Rights; former Senior Trustworthy AI Fellow Mozilla. "Closing the AI Accountability Gap" framework. URL: https://rajiinio.github.io/

Industry / former regulator verified#

1. 25. Miriam Vogel — President & CEO EqualAI. Former Chair NAIAC. Co-author: "Governing the Machine" (Bloomsbury Oct 2025). Former Associate Deputy Attorney General. URL: https://www.equalai.org/about-us/leadership/miriam-vogel/

1. 26. Reggie Townsend — VP SAS Data Ethics Practice; NAIAC member; EqualAI board. URL: https://www.globalcenter.ai/about/reggie-townsend (and his SAS bio)

1. 27. Christina Montgomery — Recently moved from IBM (VP & Chief Privacy & Trust Officer, chaired IBM AI Ethics Board) to General Motors as Chief Privacy & Trust Officer. NAIAC member. JD Harvard. URL: https://news.bloomberglaw.com/business-and-practice/gm-hires-ibms-top-ai-governance-lawyer-as-ex-law-head-retires

1. 28. Brad Smith — Microsoft Vice Chair and President. Drove Microsoft 5-point AI governance blueprint. URL: https://www.microsoft.com/en-us/about/leadership/brad-smith/

1. 29. Rumman Chowdhury — CEO Humane Intelligence; US Science Envoy for AI; former Director ML Ethics Twitter; Responsible AI Fellow Berkman Klein. URL: https://www.rummanchowdhury.com/

1. 30. Mark Surman — President Mozilla Foundation. URL: https://www.mozilla.org/en-US/about/leadership/

1. 31. Mihaela Vorvoreanu — Microsoft Research, Director UX Research and Responsible AI Education for Aether. URL: https://www.microsoft.com/en-us/research/people/mivorvor/

1. 32. Michael Rubin — Latham & Watkins Partner, Global Chair of AI; Global Vice Chair of Technology Industry Group. URL: https://www.lw.com/en/practices/artificial-intelligence

1. 33. Ilana Golbin Blumenfeld — PwC Director, Digital Assurance & Tech (Responsible AI lead). Listed "100 Brilliant Women in AI Ethics" 2020. URL: https://www.pwc.com/

Key facts not used (low credibility / could not verify)#

• User claim "IAPP 93% use / 7% governance" — could not verify exact wording in IAPP source.

Closest verified: IAPP "AI Governance Profession Report 2025": "1.5% of orgs believe they have adequate governance headcount, 77% are actively working on AI governance (rising to ~90% for orgs already using AI)"

And IAPP "AI governance in the agentic era": "99% of enterprise AI developers exploring or developing AI agents" (IBM/Morning Consult)

Honest framing: deployment/governance gap is real and supported, but the exact 93/7 fraction is not the published number — should be replaced with verified citations.