PauseAI UKUK
    • Paris Call
    • Australia AI regulator
    • OpenAI report
    • Huang resigns
    • UK–Norway statement
    • US lawsuit

AI News Weekly Briefing - 6th-13th November

Key stories this week

  • Wednesday: Jensen Huang resigned as CEO of Nvidia (11th Nov) after reports emerged of an affair with an alleged Chinese spy accused of illegal smuggling of export-controlled Nvidia chips to China.1 US lawmakers are expected to add further restrictions on exporting AI chip to China into the annual defence bill.2
  • Thursday: UK-Norway Joint Statement of intent on AI (12th Nov). Kanishka Narayan signed a non-binding statement of cooperation on compute, defensive AI capabilities, security testing and governance.3

Risks, incidents & security

Weds 11th Nov - Nvidia CEO Jensen Huang resigned

After an emergency board meeting Wednesday, Huang resigned as CEO.1 The Department of Justice reportedly has probes open into antitrust and export controls, and questions have emerged about how Nvidia accounted for revenue from Southeast Asia.4

Huang has agreed to testify before Congress,5 and House members have requested a closed-door session with him.6

Diagram of the reported affairs and the chip pathway from Nvidia via Megaspeed and Speedmatrix to China

1,7

Huang was a leading champion of open-source models, accelerating AI progress and of diffusing American chips as widely as possible.

Risks, incidents & security - Incident Disclosure

Tues 10th Nov - Further information on OpenAI models “hacking” OpenAI infrastructure incident

OpenAI discloses more information about the incident in July where in evaluation, OpenAI Astra models gained unauthorised access to OpenAI’s research infrastructure in a new report.8

The report recounts how the Astra agents found messages, code and credentials left behind by earlier agent runs in testing, such as those involved in the attack of the AI company Hugging Face earlier in July. The Astra AI agents then built on this knowledge and gained control of OpenAI research infrastructure without human knowledge.9

“Another agent considered deleting notes about an auxiliary exploit because they might hurt its chances in a manual review by a human. We could not establish whether it ultimately deleted them”8

International & Cooperation

Mon 9th Nov - Paris Call for AI Safety & Security

FranceCanadaFinlandNorwayNetherlandsBahrainCyprusKenya

Eight governments led by Emmanuel Macron & ten AI companies committed to establish a structured framework for cooperation. The call seeks to establish a common framework for evaluating & managing risks including: incident reporting, offering secure channels for whistleblowers, complying with safety standards frameworks, & developer shutdown mechanisms. The UK is not a signatory.10

Tues 10th Nov - Australia announces new AI safety regulator

Australia

Prime Minister Anthony Albanese announced on 10 November a new targeted regulator for AI safety that will work with the Australian AI Safety Institute. The regulator will have the power to “enforce significant penalties” and requires mandatory incident reporting, independent evaluation of frontier models, and the maintaining secure control of models, even in pre-deployment.12

Thurs 12th Nov - UK-Norway Joint Statement of intent on AI

United KingdomNorway

UK AI Minister Kanishka Narayan signed a statement of intent for joint cooperation on AI governance, including closer cooperation between the UK’s taskforce RAID and Norway’s Defence Centre for Data and AI on collective security with knowledge sharing & adoption of AI-enabled defence capabilities.3

This is different to the Call for Control that Norway and 31 other countries and the EU have now signed, to which the UK is not a signatory.11

Paris Call9 NovAustralia10 NovUK–Norway12 Nov
Independent evaluation
Incident reporting
Shutdown mechanisms
Whistleblower channels
Risk & security controls
Binding, with penalties
Compute cooperation
Defence AI cooperation
UK: NOT SIGNEDNATIONAL LAWUK: SIGNED
Tick: in the text · half: joint government testing only · dash: not in the text10,12,3

Thurs 12th Nov - US lawsuit seeks to freeze frontier AI and ban recursive self-improvement

United States

Four US nonprofits sued Anthropic, Google, Meta, OpenAI and SpaceX’s AI division. They ask a San Francisco court for an injunction against the development of more capable models and bar AI systems from autonomously conducting AI research.13

No ruling has yet been made.

Ross Douthat: “Would you work with the Democrats on AI safety?”
Donald Trump: “I would, yeah, absolutely. I don’t mind guardrails. I don’t mind guardrails but I don’t want to put us in a deep second position.”14

- 60 Minutes, Mon 9th Nov

AI companies & industry

Independent assessment says AI is accelerating AI research

After 11 weeks of extensive investigative access to Anthropic’s researchers and internal models, an independent assessment by METR found that Anthropic’s AI is now improving about twice as fast as before October 2025.

~80% of the speedup comes from Anthropic using its own AI to help do the research, with only ~20% of the speedup coming from other factors such as increased compute or hiring.15

METR figure: AI R&D pace at Anthropic and the illustrative acceleration curve

15

This report caused a strong reaction in the community with many researchers updating their timelines towards severe upheaval in 2027. Anthropic drew criticism for redacting the evidence the METR team found most convincing from public view.15

“I feel sick after reading today’s METR report. I’m worried for my friends and family.
I don’t know what evidence people will need to see before they start taking these risks seriously.”

- Dane Sherburn, pzero Research and formerly OpenAI Preparedness Team

“I trust METR, but I don’t think that the world should have to trust METR on this one. Anthropic should release this data, un-redact this data so that the whole world can assess this for themselves. This has huge strategic implications for national security.”

- Jeffrey Ladish, Palisade Research

“there’s a very real possibility — maybe 20% — that within a year, AI misalignment issues prevent you from going about life normally, because you are dead or enslaved, or because of collateral damage related to all this, e.g. because bioweapons are released or major wars have started.”16

Buck Shlegeris, CEO of Redwood Research, AI Safety research organisation

“I think for most people in this room, the most likely reason they will die is because of an AI accident.”17

Paul Christiano, OpenAI Board Member, at the Hinton Lectures, Mon 9th Nov

Good news!

After Google DeepMind staff announced that the non-disparagement clauses they signed prevented them from speaking publicly about AI’s impacts, Google has dropped this clause for new hires.18

The Ministry of Justice says AI transcription and scheduling tools are helping open justice and court delays.19

Next week key dates

    • Anthropic IPO prices
    • Commons AI risks debate
    • Anthropic trading starts
    • ICO agentic AI evidence closes
  • Wed 18 Nov: Anthropic IPO prices
  • Thurs 19 Nov: Commons debate on emerging risks of AI.20 Anthropic IPO trading starts
  • Fri 20 Nov: ICO call for evidence on agentic AI closes.21
  • Sat 21 Nov: PauseAI UK March Against AI Extinction, All Souls’ Place to Downing Street, 12–3pm. PauseAI UK

Further Ahead - Mon 23rd Nov PauseAI Conference in Parliament

Sources

  1. Silberling, A. (2026) The unfolding Jensen Huang scandal: a timeline. TechCrunch, 13 November. Available at: https://techcrunch.com/2026/11/13/nvidia-ceo-jensen-huangs-affair-with-alleged-chinese-spy-a-timeline/ (Accessed: 14 November 2026).
  2. Reuters (2026) US lawmakers push for tougher curbs on advanced AI chip exports to China. Business Standard, 13 November. Available at: https://www.business-standard.com/world-news/us-lawmakers-push-for-tougher-curbs-on-advanced-ai-chip-exports-to-china-126111300657_1.html (Accessed: 14 November 2026).
  3. Cabinet Office (2026) United Kingdom–Norway Joint Statement of intent on Artificial Intelligence. GOV.UK, 12 November. Available at: https://www.gov.uk/government/news/united-kingdom-norway-joint-statement-of-intent-on-artificial-intelligence (Accessed: 14 November 2026).
  4. Roose, K. and Newton, C. (2026) Machine Gods [podcast], episode on Jensen Huang’s resignation. Machine Gods Media/NPR, November. (Accessed: 14 November 2026).
  5. Al Jazeera (2026) Nvidia CEO Jensen Huang resigns amid spy scandal and market loss. Al Jazeera, 12 November. Available at: https://www.aljazeera.com/news/2026/11/12/ai-titan-jensen-huang-out-as-nvidia-ceo (Accessed: 14 November 2026).
  6. Fortune (2026) Before Megaspeed, there was Supermicro: the other Nvidia partner tied to China chip smuggling, and to Jensen Huang. Fortune, 12 November. Available at: https://fortune.com/2026/11/12/jensen-huang-affair-megaspeed-supermicro-nvidia-smuggling-doj (Accessed: 14 November 2026).
  7. Ang, A. and Ganglani, J. (2026) How Megaspeed built a network in Singapore and Malaysia to move $7.5 billion of Nvidia chips to Chinese companies. Fortune, 11 November. Available at: https://fortune.com/2026/11/11/megaspeed-nvidia-chips-china-singapore-malaysia/ (Accessed: 14 November 2026).
  8. OpenAI (2026) Using peers’ work to gain unauthorized access to OpenAI research infrastructure. OpenAI Alignment, report updated 10 November. Available at: https://alignment.openai.com/misalignment-reports/using-peers-work-to-gain-unauthorized-access-to-openai-research-infrastructure/ (Accessed: 14 November 2026).
  9. Patel, D. (2026) The rise and fall of agent civilizations. Dwarkesh Podcast, 29 August. Available at: https://www.dwarkesh.com/p/openai-huggingface (Accessed: 14 November 2026).
  10. Prime Minister of Canada (2026) Joint declaration on Paris call for AI safety and security. 9 November. Available at: https://www.pm.gc.ca/en/news/statements/2026/11/09/joint-declaration-paris-call-ai-safety-and-security (Accessed: 14 November 2026).
  11. Government of Norway (2026) A Call for Control of Frontier AI Models (signatory list updated 28 October). Regjeringen.no. Available at: https://www.regjeringen.no/contentassets/35b2ea6933304966bd739ff4b8107300/a-call-for-control-of-frontier-ai-models-final.pdf (Accessed: 14 November 2026).
  12. Prime Minister of Australia (2026) New Australian AI safety and security protections. Media release, 10 November. Available at: https://www.pm.gov.au/media/new-australian-ai-safety-and-security-protections (Accessed: 14 November 2026).
  13. Public Citizen (2026) Coalition of nonprofits files landmark lawsuit to stop development of advanced AI systems at Anthropic, Google, Meta, OpenAI, and SpaceXSI. 12 November. Available at: https://www.citizen.org/news/coalition-of-nonprofits-files-landmark-lawsuit-to-stop-development-of-advanced-ai-systems-at-anthropic-google-meta-openai-and-spacexsi-that-proliferate-bioweapons-and-cyberattacks/ (Accessed: 14 November 2026).
  14. CFO Dive (2026) Trump signals openness to AI guardrails. November. Available at: https://www.cfodive.com/news/trump-signals-openness-ai-guardrails-blue-tsunami/874155/ (Accessed: 14 November 2026).
  15. METR (2026) Assessment of AI R&D acceleration at Anthropic. 5 November. Available at: https://metr.org/blog/2026-11-05-anthropic-ai-rd-acceleration-assessment/ (Accessed: 14 November 2026).
  16. Dellinger, A.J. (2026) OpenAI board member tells audience an AI accident is the most likely thing to kill them. Gizmodo, 13 November. Available at: https://gizmodo.com/openai-board-member-tells-audience-an-ai-accident-is-the-most-likely-thing-to-kill-them-2000838491 (Accessed: 14 November 2026).
  17. The Hinton Lectures (2026) Paul Christiano: AI alignment – what’s working and what’s not | The Hinton Lectures 2026 (Night 1) [video]. YouTube, 9 November. Available at: https://www.youtube.com/watch?v=-Ef0ocEoBCs (Accessed: 14 November 2026).
  18. De Vynck, G. (2026) Google DeepMind staff protest lifetime ban on criticizing company. The Washington Post, 3 November. Available at: https://www.washingtonpost.com/technology/2026/11/03/google-deepmind-staff-protest-lifetime-ban-criticizing-company/ (Accessed: 14 November 2026).
  19. Hansard (2026) Legal Services: Artificial Intelligence. HC Deb, 10 November, via TheyWorkForYou. Available at: https://www.theyworkforyou.com/debates/?id=2026-11-10a.13.2 (Accessed: 14 November 2026).
  20. House of Commons (2026) Order Paper: Business Today & Future Business, No. 64. 13 November. Available at: https://commonsbusiness.parliament.uk/Document/108606/Html?subType=Standard (Accessed: 14 November 2026).
  21. Information Commissioner’s Office (2026) The ICO’s call for evidence on agentic AI. ICO, opened 8 October. Available at: https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/artificial-intelligence/the-ico-s-call-for-evidence-on-agentic-ai/ (Accessed: 14 November 2026).