Anthropic, OpenAI and Google have been designing their own AI standards body since July
Working groups from Anthropic, OpenAI and Google, staffed by executives below chief-executive level, have met regularly since July on a voluntary industry body that would set safety-testing and auditing rules, The Information reported on 13 September, citing people familiar with the talks. The discussions began before Dario Amodei's weekend call for coordination and, in Seoul Economic Daily's reading, aim at a self-regulatory framework while a draft White House order for a standards body sits stalled. None of the three companies has confirmed the report.
Artificial Intelligence··Midday
Talks since July, before Amodei's call
The Information reported on 13 September, citing people familiar with the matter, that Anthropic, OpenAI and Google have for some time been talking among themselves about creating a standards body for the AI industry that would set safety benchmarks and other rules. The working groups are made up of executives below chief-executive level, have met regularly since July and aim to build a self-regulatory framework without government involvement. The talks began before Anthropic chief executive Dario Amodei called on Saturday for companies to coordinate on testing and auditing, and Seoul Economic Daily writes that the teams have continued them for months.[1], [2]
Altman: the labs will have to build it themselves
OpenAI chief executive Sam Altman told a companywide meeting last week that he supported a testing and auditing organisation for the industry but believed the major laboratories would have to create it on their own, without the support of the US government. Amodei, for his part, stressed the need for independent oversight on 12 September while arguing for a slower pace of development. Seoul Economic Daily notes that Anthropic and OpenAI, regarded as each other's biggest rivals, had traded public attacks ahead of their initial public offerings and now appear to have concluded that setting up the voluntary body takes priority over the feud.[1], [2]
A stalled White House draft, and what the report leaves open
The industry effort runs alongside a government track that has not moved: a draft executive order prepared at the White House for a new AI standards body ran into internal conflict, and David Sacks, co-chair of the President's Council of Advisors on Science and Technology, opposes a government-led body. Seoul Economic Daily reads the companies' pursuit of a voluntary organisation as an attempt to launch a private-sector-led body after those discussions failed to progress, at a moment when calls to slow AI development are gaining traction. The stage the talks have reached, the body's name and the powers it would hold were not disclosed, and none of the three companies has publicly confirmed the report.[1], [2]
