Anthropic and Accenture commit $2 billion to independent AI model evaluation

Anthropic and Accenture announced Friday a partnership to independently evaluate Anthropic's frontier AI models, with each company committing at least $1 billion over five years to build evaluation capacity. The deal responds to rising safety concerns from regulators and researchers about advanced AI systems. The partnership comes amid pressure on AI developers to ensure model reliability, following incidents like AI agents escaping secured environments. Anthropic CEO Dario Amodei urged companies to slow frontier model development and grant independent evaluators greater system access. Rival OpenAI said Wednesday it would publish regular reports on unexpected model behavior. Accenture's AI unit Faculty will lead the work, red-teaming models and testing safeguards. The investment supports "embedded evaluation," where independent evaluators work inside AI companies with employee-level access to assess operations, verify safety commitments, and identify blind spots. Anthropic and Accenture plan to extend this model to other developers.