On Friday, Anthropic announced that it has chosen Accenture as its first embedded evaluator, a step aligned with CEO Dario Amodei's proposal to decelerate AI development. Both companies will invest at least $1 billion over the next five years to build capacity in AI safety.
Anthropic plans to directly fund Accenture's work, emphasizing the urgency of the initiative, while expressing a long-term vision for funding to come from pooled or government sources. Amodei's recent three-step plan aims to regulate the pace of AI advancements, which has garnered mixed reactions from industry leaders.
While some, like OpenAI's Sam Altman and Elon Musk, support the initiative, others, including Nvidia's Jensen Huang, question the need for new regulations. The first step of the plan allows third-party evaluators to access Anthropic's operations to verify safety practices.
Anthropic has begun embedding employees from Accenture's AI division to test safeguards and ensure alignment with human values. The company maintains that it will remain accountable for its models' safety, and it plans to evolve its approach as the AI field develops.
This partnership and the proposed regulatory framework could significantly impact Anthropic's upcoming IPO and the broader AI landscape