Anthropic partners with Accenture for independent AI safety evaluations
AI company Anthropic has agreed with consulting firm Accenture to conduct independent safety evaluations of its frontier AI models. Both companies plan to invest at least $1 billion each into the effort over the next five years.
Anthropic has announced a new partnership with consulting firm Accenture, under which Accenture will serve as a third-party evaluator assessing the safety compliance of Anthropic's frontier AI models. Under the arrangement, Accenture's evaluators will be embedded within Anthropic's operations, working closely alongside the AI company.
At least $1 billion from each side
Both companies intend to invest at least $1 billion each into this joint effort over the next five years. The partnership marks one of the first concrete steps toward a plan proposed a week earlier by Anthropic CEO Dario Amodei aimed at slowing the pace of AI development.
Amodei's proposed three-step approach calls on all "frontier AI" companies to grant ongoing, employee-like access to embedded third-party evaluators. These evaluators would be responsible for ensuring safety practices and commitments are upheld, reporting incidents, and assessing the alignment of AI models.
Access to the model development process
According to Anthropic, Accenture's evaluators will observe models as they take shape during training, follow the decisions that govern how those models are built and deployed, and speak directly with employees. Anthropic said Accenture was chosen for its experience helping businesses and governments deploy AI across multiple industries, as well as its understanding of how enterprises actually use AI in practice.
However, Anthropic acknowledged that the concept of embedded third-party evaluators is still new, and standards have not yet been established regarding the scope of evaluators' work, the level of access they receive, or the procedures for reporting concerns. As a result, Anthropic said it plans to announce additional evaluator partners in the coming weeks beyond Accenture. The company also noted that Accenture will work with other AI companies besides Anthropic.
