source: TechCrunch AI: OpenAI, Anthropic, Google have been in talks on AI safety for weeks
level: business
openai's global policy chief chris lehane said the company has been working with rivals anthropic and google deepmind on ai safety for weeks. the talks follow anthropic ceo dario amodei's essay urging the industry to slow frontier ai development to avoid catastrophic risks. lehane is in washington to discuss those risks with lawmakers. the three firms have also been creating a standards body for the ai industry, according to the information.
amodei proposed a narrow government waiver to allow safety coordination without antitrust risk, but lehane said the firms don't need one. altman hinted at private discussions last week and said openai would embed third-party evaluators. however, president trump dismissed safety concerns as a hoax, and advisor david sacks called existential risk fears overblown. lehane also backed a provision in the frontier act requiring independent verification organizations inside top labs.
the coordination comes as google's demis hassabis called in july for a u.s. standards body to screen advanced models and coordinate slowdowns. the white house has resisted, arguing any slowdown would benefit china. the talks confirm earlier reporting and show top labs moving toward shared safety practices despite regulatory uncertainty. no formal agreement has been announced, and techcrunch has reached out to the companies for comment.
why it matters: coordination among leading ai labs could set industry-wide safety standards, but it also raises antitrust concerns and clashes with current u.s. policy.
source: TechCrunch AI: OpenAI, Anthropic, Google have been in talks on AI safety for weeks