Technology
Anthropic picks Accenture as external safety evaluator in billion-dollar initiative
Anthropic is teaming up with Accenture on safety evaluations under a five-year commitment, according to CNBC.
The short version
- Anthropic announced Accenture as its first embedded evaluator to assist in executing CEO Dario Amodei's AI slowdown plan.[CNBC]
- The two firms pledged to commit a minimum of $1 billion across five years, with Anthropic directly funding Accenture's safety efforts.[CNBC]
- Staff from Accenture's AI unit Faculty will embed inside Anthropic to assess safeguards, red-team systems, and test alignment.[CNBC]
- Anthropic noted the partnership is non-exclusive and that talks are underway with other organizations, including research group METR.[CNBC]
Key facts
- Anthropic designated Accenture as an embedded evaluator to help enact its proposal to slow down artificial intelligence development.[CNBC]
- Anthropic and Accenture agreed to commit at least $1 billion over five years toward building AI safety capacity.[CNBC]
- Staff from Faculty, Accenture's AI division, will be placed inside Anthropic to red-team models and check alignment.[CNBC]
- Anthropic is directly funding Accenture's work while holding discussions with other potential third-party evaluators including METR.[CNBC]
What remains uncertain
- The precise funding structures and final agreements for future non-exclusive third-party evaluators beyond Accenture remain undetermined.[CNBC]
Sources
Outlet counts describe coverage, not independent confirmation. Reports may share a wire service or original source.