According to insiders, before Anthropic's CEO Dario Amodei called on AI companies to collaborate on technical testing and audits last Saturday, Anthropic, OpenAI, and Google had already begun discussions to establish an AI industry standards organization.

Insiders said that since July, representatives from the three companies at the level below their CEOs have formed a working group that regularly meets to discuss. Sam Altman, CEO of OpenAI, mentioned earlier this week during a company-wide meeting that he supports the establishment of an AI industry testing and audit institution, but believes that major AI labs must independently set up such a standard organization without support from the U.S. government—although he originally hoped for government support, there is currently no sufficient consensus within the industry, and some companies are opposing it.

Disagreements Over Three Models: FAA, FINRA, or Industry Self-Regulation

Amodei published a blog post titled "We Must Slow Down the Pace of Frontier Development" last Saturday, advocating for slowing down AI development and establishing an independent oversight mechanism, with industry self-regulation being a key part of it. He proposed that while the government begins to formulate AI regulatory laws, companies can first implement voluntary safety standards, and the government could take the lead or mediate industry dialogue; he also called on leading global companies to grant external evaluators like METR access to technology equivalent to that of internal employees, and pledged that Anthropic would be the first to implement this. Additionally, he urged limited antitrust exemptions to enable safe collaboration.

Demis Hassabis, who recently stepped down as CEO of Google DeepMind, said the proposal was in the right direction but some details "still need refinement"—he had previously suggested building an AI self-regulatory organization based on the model of the U.S. Financial Industry Regulatory Authority (FINRA) in July. Elon Musk, CEO of SpaceX, also expressed his support. According to a previous report by The Information, the White House had circulated a draft executive order proposing the establishment of this regulatory body, but failed to gain enough support within the government by late summer, and the related work was put on hold; David Sacks, former White House AI director, has openly opposed the establishment of an institution similar to FINRA.

Criticism: This Is "a Cartel's Design"

There are also many criticisms within the industry. Aiden Gomez, CEO of Cohere, expressed concern about large AI companies jointly setting rules to restrict smaller competitors: "This idea sounds exactly like a cartel's design." Sacks questioned whether slowing down development was entirely altruistic, arguing that product liability risks themselves provide economic incentives to prevent harm.

According to insiders, the working group meetings between the three companies regarding the industry-led standard organization are still ongoing, with meetings held just last week. It remains unclear about Meta's position; its Chief AI Officer Alexander Wang only stated on the X platform that "alignment is the foundation of building a personal superintelligence for all people," without clearly supporting or opposing Amodei's entire proposal.