Menu Close

Anthropic, OpenAI, and Google discuss industry AI safety standards body

Frontier AI standards and testing board with Anthropic, OpenAI, and Google chips, TALKS ONLY and NO BODY LAUNCHED stamps.

The Washington Post and CNN report frontier labs have been talking about a shared standards and testing organization, talks that began before Amodei’s weekend pacing essay.

Anthropic, OpenAI, and Google have discussed creating an industry AI safety standards and testing body, The Washington Post reported on September 14. CNN, citing people familiar with the situation, corroborated that the companies have talked about a shared standards organization amid rising safety concerns.

According to CNN’s account, the catalyst was a July essay by Google DeepMind founder Demis Hassabis proposing a U.S.-led Frontier AI Standards Body modeled in part on FINRA, the brokerage industry’s self-regulatory organization. Hassabis argued for testing advanced models before deployment through a public-private partnership overseen by government, funded by industry, and staffed by independent technical experts. People familiar with the talks told CNN representatives have participated in working-group discussions since July. The White House has separately described ongoing work with industry on a voluntary pre-release review framework, CNN reported.

The Washington Post framed the discussions alongside public calls from leaders at the three companies to slow frenetic frontier competition. That package is related to, but distinct from, Anthropic CEO Dario Amodei’s September 13 “pace the frontier” essay and interview tour already on AI Tech Daily: CoS notes the standards-body talks predate that weekend essay. No formal body has been announced, and reports do not establish that a charter, membership, or mandatory testing regime has been agreed.

This brief covers reported discussions only. It does not invent a launched regulator or claim the companies have paused capability development.

Sources

0 0 votes
Article Rating
Subscribe
Notify of
0 Comments
Inline Feedbacks
View all comments
0
Would love your thoughts, please comment.x
()
x