OpenAI, Anthropic, and Google Collaborate on AI Safety
nashnova research
OpenAI confirmed it has been coordinating with Anthropic and Google DeepMind on AI safety for weeks, while all three push federal safety legislation in Washington — but with Congress about to recess and the president openly opposing AI restrictions, whether the effort produces real standards remains an open question.
Three rivals joining hands on safety — what exactly are they doing?
OpenAI's global policy chief Chris Lehane confirmed Tuesday at a Washington briefing that the three companies have been in safety discussions for weeks.
Lehane stressed the coordination requires no antitrust waiver, citing the airline industry — where competitors routinely share safety data — as precedent.
This means → the three chose a "just start, don't wait for legal permission" path, sidestepping the biggest barrier to cooperation: antitrust review.
Why now? Who spoke first?
The immediate trigger: Anthropic CEO Dario Amodei published a ~3,800-word essay last Saturday calling for a slowdown in frontier AI development so researchers can better assess potential threats.
OpenAI CEO Sam Altman and xAI head Elon Musk quickly endorsed the idea publicly.
In plain terms = these three are day-to-day competitors, yet on the question of "AI could go seriously wrong," they have taken the rare step of standing on the same side.
What legislation is moving in Washington?
House bill: Republican Rep. Jay Obernolte and Democratic Rep. Lori Trahan jointly proposed a federal AI governance framework.
Senate bill: Majority Leader John Thune, Republican Sen. Ted Cruz, and Democratic Sen. Amy Klobuchar are leading a bill focused on catastrophic AI risk.
Antitrust-waiver bill: Republican Sen. Jim Banks and Democratic Sen. Adam Schiff proposed limited antitrust immunity so AI companies can share information on loss-of-control risks, cyber/bio threats, and data theft by Chinese firms. Sources say it has been folded into the annual defense authorization bill.
Can any of these bills actually pass? Where are the obstacles?
The House recesses this Thursday and will not return until after the November election; the post-election lame-duck session is expected to focus on the Pentagon budget, leaving little room for AI legislation.
The bigger headwind comes from the White House: President Trump posted multiple times on social media Monday, forcefully opposing any AI restrictions and calling existential-risk concerns a "hoax."
This means → the industry side has shaken hands, but the policy window is closing — whether this cooperation becomes real industry standards depends on the post-election political landscape.
市场有风险，内容仅供研究参考，不构成投资建议。