Anthropic taps Accenture as first safety evaluator — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic selected Accenture as its first embedded evaluator on Friday, the company's first concrete step toward implementing CEO Dario Amodei's Saturday proposal to slow the pace of AI development.
- Both Anthropic and Accenture agreed to invest at least $1 billion over five years to build evaluator capacity, though Anthropic will fund Accenture's work directly given the urgency.
- Anthropic said long-term funding should come from pooled or government sources, citing its June Advanced AI Framework, but acknowledged neither exists today and plans to work with evaluators under different funding arrangements.
- Amodei's three-step plan drew split reactions: OpenAI CEO Sam Altman and Elon Musk cheered it, while Nvidia CEO Jensen Huang brushed off concerns and argued against new regulation.
- Anthropic will initially embed employees from Faculty, Accenture's specialist AI business, to test safeguards, red-team models, and assess alignment with human values; the partnership is non-exclusive, with discussions underway with research nonprofit METR.
- Anthropic emphasized it retains full responsibility for model safety and that working with embedded evaluators does not reduce its accountability.
Why it matters: Anthropic is moving AI safety oversight from rhetoric to operational infrastructure — committing $1B+ and granting employee-level third-party access before any government mandate exists. If the embedded-evaluator model holds, competitors face pressure to match that transparency or explain why they won't.
Ask SkimNews
