OpenAI Confirms AI Safety Talks With Anthropic, Google — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI disclosed it has been in safety talks with Anthropic and Google DeepMind for weeks, with global policy chief Chris Lehane confirming the coordination while meeting with reporters in Washington on Tuesday.
- Anthropic CEO Dario Amodei's Saturday essay called on the industry to slow frontier AI and avoid catastrophic risks, gaining public support from OpenAI's Sam Altman, Google's Demis Hassabis, and Elon Musk.
- Altman said OpenAI would join Anthropic in embedding third-party evaluators to monitor for safety, while warning that the coordination could risk antitrust violations if it were found to suppress competition.
- The three labs are working together to create an AI standards body, per The Information — and Altman told staff this body would need to form without U.S. government support.
- Lehane said the firms don't need the narrow government waiver Amodei proposed for safety coordination, but OpenAI separately backs a FRONTIER Act provision that would require frontier labs to admit "independent verification organizations" inside.
- Hassabis called in July for a U.S.-established standards body with the power to screen advanced models and coordinate industrywide slowdowns, while President Trump has dismissed safety concerns as a hoax and pushed back on tighter regulation.
Why it matters: Three frontier labs coordinating on safety — a move Altman himself flagged could violate antitrust law — puts Washington under pressure to either bless the arrangement via legislation like the FRONTIER Act or watch the standards body form without it, as Altman told staff. Trump's dismissal of the risks deepens the gap between the industry and the administration.
Ask SkimNews
