Skip to main content

Tech Giants Quietly Discuss AI Safety Standards Group

Secret Talks Among AI Giants

According to insiders, Anthropic, OpenAI, and Google have been quietly discussing the formation of an AI industry standards organization since July. The talks involve representatives just below the CEO level, who have formed a working group that meets regularly. This revelation comes on the heels of Anthropic CEO Dario Amodei's public call for collaboration on technical testing and audits.

OpenAI CEO Sam Altman reportedly voiced support for an industry-led testing and audit body during a company-wide meeting this week. However, he believes major AI labs must establish such standards independently, without relying on U.S. government support. While Altman initially hoped for government backing, there's currently insufficient consensus in the industry, with some companies opposing the idea.

Three Models, No Agreement

Amodei published a blog post last Saturday titled "We Must Slow Down the Pace of Frontier Development," advocating for a slower AI development pace and an independent oversight mechanism. He proposed that companies adopt voluntary safety standards while governments work on AI regulations. He also urged leading global companies to grant external evaluators like METR access equivalent to internal employees, pledging Anthropic would lead by example. Additionally, he called for limited antitrust exemptions to enable safe collaboration.

Demis Hassabis, who recently stepped down as CEO of Google DeepMind, said the proposal is directionally correct but needs refinement. He had previously suggested building an AI self-regulatory organization modeled after the U.S. Financial Industry Regulatory Authority (FINRA) in July. Elon Musk, CEO of SpaceX, also expressed support.

A previous report by The Information revealed that the White House had circulated a draft executive order to establish such a regulatory body, but it failed to gain enough support by late summer, and the effort stalled. David Sacks, former White House AI director, has openly opposed creating an institution similar to FINRA.

"A Cartel's Design"?

Critics within the industry are vocal. Aiden Gomez, CEO of Cohere, expressed concern that large AI companies could set rules to restrict smaller competitors: "This idea sounds exactly like a cartel's design." Sacks questioned whether slowing development was entirely altruistic, arguing that product liability risks already provide economic incentives to prevent harm.

Meanwhile, the working group meetings among the three companies are ongoing, with a meeting held just last week. Meta's position remains unclear; its Chief AI Officer Alexander Wang only stated on X that "alignment is the foundation of building a personal superintelligence for all people," without clearly supporting or opposing Amodei's full proposal.

Key Points

  • Secret discussions: Anthropic, OpenAI, and Google have been meeting since July to form an AI standards group.
  • Altman's stance: Supports industry-led testing but insists on independence from government.
  • Amodei's proposal: Advocates slowing AI development and voluntary safety standards.
  • Criticism: Cohere CEO calls it "a cartel's design"; others question motives.
  • Meta's silence: No clear position on the proposal.
  • White House effort: A draft executive order failed to gain traction.