AI Labs' SAFA Watchdog Eyes Rice, Krishnan for Top Roles
By @sharedot · · 6 pages
- AI
- AI Regulation
- AI Safety
- Self Regulation
The lab-backed AI standards body now has a name, SAFA, and Stocktwits reports high-profile candidates including Condoleezza Rice are shortlisted for its leadership.
SAFA: the watchdog gets a name and a launch window
The industry-led AI safety body reportedly under discussion by Google, OpenAI and Anthropic now has a tentative identity: the Standards Authority for Frontier AI, or SAFA.
According to Stocktwits, citing The Information, the three companies hope to launch the organization by the end of 2026 or early 2027.
SAFA would set standards for testing and reporting AI safety issues, support third-party pre-deployment model testing, define how voluntary lab safety commitments work in practice, and establish qualifications for independent auditors.
The group's existence remains unconfirmed by the companies themselves — Stocktwits says Google, OpenAI and Anthropic did not respond to requests for comment.
A shortlist of Washington heavyweights
Stocktwits reports that the working group has weighed prominent figures for leadership.
Sriram Krishnan, a former venture capitalist and top AI policy adviser in the Trump administration, and Arati Prabhakar, former Biden-era director of the White House Office of Science and Technology Policy, were reportedly considered for the CEO role.
Former Secretary of State Condoleezza Rice and entrepreneur David Friedberg were reportedly on a shortlist for chair.
Potential scientific advisers include METR founder Beth Barnes and Paul Christiano, a Commerce Department adviser and recent OpenAI foundation board appointee — a roster that signals an attempt at bipartisan, technically credible leadership.
Why it is surprising: no government backing
The surprising turn is the governance plan itself.
FindArticles reports that the talks began as early as July as a working group and originally centered on a public-private partnership proposed by Google DeepMind's Demis Hassabis, under which industry funding and independent experts would evaluate advanced models before release.
According to Stocktwits, that public-private model has stalled: the Trump administration told the companies they needed industry consensus before government involvement, so SAFA is now planned to operate independently, though some representatives remain optimistic about a future government relationship.
Stakes: credibility versus concentrated power
FindArticles reports that OpenAI CEO Sam Altman favored an industry testing and auditing body and believed leading labs might need to create it without government backing — while noting no rulebook, membership or enforcement mechanism has been publicly agreed, and that testing and auditing are distinct functions any body would need to separate.
The outlet also reports Cohere CEO Aidan Gomez criticized the prospect of the largest firms setting the rules, arguing it could disadvantage smaller rivals, and that House Speaker Mike Johnson had found no consensus on AI guardrails.
Working-group members are reportedly debating whether SAFA should run its own testing alongside the Commerce Department's CAISI, which some believe lacks resources. Whether SAFA earns trust or entrenches incumbents will hinge on those unanswered details.