Anthropic has selected Accenture as a partner to conduct independent evaluations of its AI systems. Accenture specialists will be integrated into the laboratory's internal workflows to monitor safety compliance, perform stress testing, and assess model alignment with human values.
The companies plan to invest at least $1 billion over the next five years to build capacity in this field. Anthropic will provide direct funding for Accenture's work for the time being, as the company seeks to establish future joint or government-backed funding sources.
This partnership is part of a three-stage plan proposed by Anthropic CEO Dario Amodei to slow the development of advanced AI models. The initiative has received support from OpenAI CEO Sam Altman and xAI leader Elon Musk, while Nvidia CEO Jensen Huang expressed skepticism regarding the need for additional regulation.
The agreement with Accenture is not exclusive. Anthropic is currently in negotiations with the non-profit research organization METR and other potential partners to expand its evaluation capabilities. The company intends to share updates as its assessment practices evolve and new evaluators are brought on board.
