Anthropic, OpenAI, Google in Secret Talks to Form AI Safety Standards Body
According to industry sources, prior to Anthropic CEO Dario Amodei’s call last Saturday for AI firms to collaborate on technical testing and audits, Anthropic, OpenAI, and Google had already initiated discussions to form an AI industry standards body.
Insiders report that since July, representatives from these three companies, operating below the CEO level, have formed a working group that meets regularly. Earlier this week, OpenAI CEO Sam Altman stated during an all-hands meeting that he supports creating an independent AI testing and audit institution, arguing that major labs must establish such standards without U.S. government involvement. Although he initially favored government backing, there is currently insufficient industry consensus, with some companies opposing the move.
Debate Over Three Regulatory Models: FAA, FINRA, or Self-Regulation
Last Saturday, Amodei published a blog post titled "We Must Slow Down the Pace of Frontier Development," urging a slowdown in AI progress and the creation of an independent oversight mechanism, with industry self-regulation as a core component. He suggested that while governments draft AI regulations, companies should first adopt voluntary safety standards, with the government potentially leading or mediating industry dialogue. He also called on leading global firms to grant external evaluators, such as METR, access to technology equivalent to that of internal employees, pledging that Anthropic would be the first to do so. Additionally, he advocated for limited antitrust exemptions to facilitate safe collaboration.
Demis Hassabis, who recently stepped down as CEO of Google DeepMind, described the proposal as a step in the right direction but noted that certain details "still need refinement." In July, he had suggested creating an AI self-regulatory organization modeled after the U.S. Financial Industry Regulatory Authority (FINRA). Elon Musk, CEO of SpaceX, also voiced support. According to a previous report by The Information, the White House circulated a draft executive order proposing this regulatory body, but failed to secure sufficient government support by late summer, halting related efforts. David Sacks, the former White House AI director, has openly opposed establishing an institution similar to FINRA.
Backlash: Critics Label It "a Cartel's Design"
The proposal has faced significant criticism within the industry. Aiden Gomez, CEO of Cohere, expressed concern that large AI companies might jointly set rules to restrict smaller competitors, stating, "This idea sounds exactly like a cartel's design." Sacks questioned whether slowing development was purely altruistic, arguing that product liability risks already provide economic incentives to prevent harm.
According to insiders, working group meetings between the three companies regarding this industry-led standard organization are ongoing, with sessions held as recently as last week. Meta’s position remains unclear; its Chief AI Officer, Alexander Wang, posted on X that "alignment is the foundation of building a personal superintelligence for all people," without explicitly supporting or opposing Amodei’s broader proposal.
Related article
OpenAI Launches Task Force to Probe Unusual Codex Usage
OpenAI has launched a dedicated investigation team following user reports of irregular Codex quota usage. Paid ChatGPT subscribers report a sharp decline in weekly quotas, with balances dropping from 96% to zero rapidly, even during periods of minima
Societe Generale bets on AI to slash costs by up to 600 million euros
Bloomberg reports that Société Générale SA anticipates artificial intelligence’s expanding role will significantly drive cost reduction initiatives.The Paris-based institution estimates current AI-driven savings potential between €500 million and €60
How to create free AI talking avatars
Master Roblox Avatar Customization: A Guide Inspired by Karl JacobsIntroductionEnhancing Your AvatarNavigating the Avatar EditorBrowsing Trending LooksTailoring Your CharacterSearch and Filter ToolsLeveraging the Search FunctionFiltering by Popularit
Related Special Topic Recommendations
Comments (0)
0/500
According to industry sources, prior to Anthropic CEO Dario Amodei’s call last Saturday for AI firms to collaborate on technical testing and audits, Anthropic, OpenAI, and Google had already initiated discussions to form an AI industry standards body.
Insiders report that since July, representatives from these three companies, operating below the CEO level, have formed a working group that meets regularly. Earlier this week, OpenAI CEO Sam Altman stated during an all-hands meeting that he supports creating an independent AI testing and audit institution, arguing that major labs must establish such standards without U.S. government involvement. Although he initially favored government backing, there is currently insufficient industry consensus, with some companies opposing the move.
Debate Over Three Regulatory Models: FAA, FINRA, or Self-Regulation
Last Saturday, Amodei published a blog post titled "We Must Slow Down the Pace of Frontier Development," urging a slowdown in AI progress and the creation of an independent oversight mechanism, with industry self-regulation as a core component. He suggested that while governments draft AI regulations, companies should first adopt voluntary safety standards, with the government potentially leading or mediating industry dialogue. He also called on leading global firms to grant external evaluators, such as METR, access to technology equivalent to that of internal employees, pledging that Anthropic would be the first to do so. Additionally, he advocated for limited antitrust exemptions to facilitate safe collaboration.
Demis Hassabis, who recently stepped down as CEO of Google DeepMind, described the proposal as a step in the right direction but noted that certain details "still need refinement." In July, he had suggested creating an AI self-regulatory organization modeled after the U.S. Financial Industry Regulatory Authority (FINRA). Elon Musk, CEO of SpaceX, also voiced support. According to a previous report by The Information, the White House circulated a draft executive order proposing this regulatory body, but failed to secure sufficient government support by late summer, halting related efforts. David Sacks, the former White House AI director, has openly opposed establishing an institution similar to FINRA.
Backlash: Critics Label It "a Cartel's Design"
The proposal has faced significant criticism within the industry. Aiden Gomez, CEO of Cohere, expressed concern that large AI companies might jointly set rules to restrict smaller competitors, stating, "This idea sounds exactly like a cartel's design." Sacks questioned whether slowing development was purely altruistic, arguing that product liability risks already provide economic incentives to prevent harm.
According to insiders, working group meetings between the three companies regarding this industry-led standard organization are ongoing, with sessions held as recently as last week. Meta’s position remains unclear; its Chief AI Officer, Alexander Wang, posted on X that "alignment is the foundation of building a personal superintelligence for all people," without explicitly supporting or opposing Amodei’s broader proposal.
OpenAI Launches Task Force to Probe Unusual Codex Usage
OpenAI has launched a dedicated investigation team following user reports of irregular Codex quota usage. Paid ChatGPT subscribers report a sharp decline in weekly quotas, with balances dropping from 96% to zero rapidly, even during periods of minima
Societe Generale bets on AI to slash costs by up to 600 million euros
Bloomberg reports that Société Générale SA anticipates artificial intelligence’s expanding role will significantly drive cost reduction initiatives.The Paris-based institution estimates current AI-driven savings potential between €500 million and €60
How to create free AI talking avatars
Master Roblox Avatar Customization: A Guide Inspired by Karl JacobsIntroductionEnhancing Your AvatarNavigating the Avatar EditorBrowsing Trending LooksTailoring Your CharacterSearch and Filter ToolsLeveraging the Search FunctionFiltering by Popularit





Home






