Anthropic announced on Friday that it has appointed Accenture as its first embedded evaluator to assess AI safety protocols, a move tied to a broader proposal to temper the speed of artificial intelligence development. The partnership, detailed in a joint announcement, will involve Accenture’s AI division Faculty conducting red-teaming, alignment assessments and safeguard testing on Anthropic’s models. The companies have each pledged at least $1 billion over the next five years to fund the initiative, which aims to address concerns about uncontrolled recursive self-improvement in AI systems. Anthropic’s CEO Dario Amodei had previously outlined a three-step plan calling for independent evaluators with employee-level access to review AI progress, a concept he said could help prevent systems from outpacing human oversight. While OpenAI CEO Sam Altman and Elon Musk endorsed the proposal, Nvidia CEO Jensen Huang expressed skepticism about imposing regulatory constraints. The collaboration marks one of the first efforts to establish dedicated funding and expertise for independent AI evaluation, a field Amodei noted lacks established mechanisms. Accenture’s chair Julie Sweet described embedded evaluation as a critical component of the emerging safety landscape, though specific implementation details remain under development. The partnership is non-exclusive, with Anthropic indicating plans to engage additional evaluators in the coming weeks.