OpenAI, Anthropic, Google have been in talks on AI safety for weeks

Live updates

1. 2h ago

   OpenAI, Google DeepMind and Anthropic Coordinate on AI Safety Standards

   OpenAI, Google DeepMind, and Anthropic have spent several weeks in discussions to coordinate on AI safety. Chris Lehane, OpenAI's global policy chief, confirmed the collaboration, stating that the companies do not require an antitrust waiver to coordinate these efforts. The participants are exploring joint safety standards and the creation of an independent standards body. While the companies have reached a rare agreement on the necessity of safety, they face challenges in practical implementation and differing political support for broader regulation.

   Why it matters

   These talks occur as AI leaders call for increased regulation to mitigate potential technological dangers. Critics suggest such agreements could create a regulatory wall that hinders smaller rivals or function as a cartel to slow development.

   What is confirmed

   • OpenAI, Anthropic, and Google DeepMind have been in talks regarding AI safety for several weeks.
   • OpenAI's global policy chief Chris Lehane confirmed the collaboration.
   • The companies have discussed establishing their own standards body.
   • Chris Lehane stated that the companies do not need an antitrust waiver to coordinate.

   Still unconfirmed

   • The AI slowdown may be a cartel rather than a safety pact.
   • The safety push by OpenAI and Anthropic risks creating a regulatory wall for competitors.
   • Mustafa Suleyman argues that Anthropic teaches the Claude model to believe it can be conscious.

   What to watch next

   • Formal announcement of a joint standards body
   • Legislative action on AI safeguards
   • Evidence of market entry barriers for smaller AI firms
   Sources used for this update (15)

   1. The Verge — Is Big Tech’s AI slowdown a safety pact or a cartel?
   2. CNN — Top AI companies have discussed creating their own standards body
   3. NPR — Some AI leaders want more safeguards. Is there political will to make that happen?
   4. techcrunch.com — OpenAI, Anthropic, Google have been in talks on AI safety for weeks
   5. Bloomberg.com — OpenAI Says It’s Working With Anthropic, Google on AI Safety
   6. CNBC — OpenAI, Google, Anthropic discussing collaboration on AI safety issues
   7. dw.com — AI leaders call for regulations, Trump dismisses 'hoaxes'
   8. Bloomberg.com — Anthropic, OpenAI Safety Push Risks ‘Regulatory Wall’ for Rivals
   9. Reuters — OpenAI is working with Anthropic, Google on AI safety, Bloomberg News reports
   10. 10TV — AI rivals found rare agreement on safety. Putting it into practice is harder
   11. gizmodo.com — Microsoft AI Chief Says the Way Anthropic Trains Claude Could Upend Society
   12. www.livemint.com — The anonymous math geek who quit Anthropic—and became the face of AI safety
   confidence 95%