Partnering with Accenture on embedded evaluation
We're partnering with Accenture on independent evaluation of frontier AI. This is an important step toward the commitment, made in our CEO’s essay “We Must Pace the Frontier,” to embed evaluators within Anthropic.
TL;DR
- Anthropic partners with Accenture for independent evaluation of frontier AI.
- Accenture's Faculty, a specialist AI business, will lead the evaluation, including red-teaming, alignment assessments, and safeguard testing.
- Both companies plan to invest at least $1 billion each over five years in this area.
- Embedded evaluators will work inside AI companies with access comparable to employees, allowing deeper insights into model development and deployment.
- This initiative aims to make safety commitments more verifiable without reducing Anthropic's accountability.
- The framework for embedded evaluation, including data access and reporting standards, is still under development.
- Anthropic will directly fund Accenture's work initially, while exploring partnerships with other non-profit evaluators like METR.
- The partnership is non-exclusive, with plans for Anthropic to work with other evaluators and Accenture with other AI developers.