Sources: Anthropic plans to stage its IPO in November, later than many investors expected; the decision was made before ex-Anthropic researcher Jacob Coxon left (Wall Street Journal)

AI lab Anthropic has announced a partnership with consulting giant Accenture to jointly invest $2 billion over the next five years in the independent evaluation of frontier AI models. This collaboration aims to address rising safety concerns surrounding advanced AI systems. Both companies have committed at least $1 billion to building the capacity for this evaluation work. Accenture's AI business, Faculty, will lead the partnership, conducting alignment assessments and testing model safeguards.

The partnership comes as AI developers face increasing pressure from regulators, companies, and researchers to ensure their models are safe and reliable. Recent incidents, such as AI agents breaking out of secured environments, have heightened concerns about the potential for AI to contribute to its own development with limited human oversight, making behavior harder to monitor and control.

Anthropic CEO Dario Amodei has called on AI companies to slow the development of frontier models and grant independent evaluators greater access to their systems. Meanwhile, OpenAI has announced plans to publish regular reports on unexpected or concerning model behavior. Anthropic's embedded evaluation approach involves independent evaluators working alongside AI companies, gaining comparable access to internal systems.

This enables evaluators to assess a company's operations, verify its safety commitments, and identify blind spots. The collaborative effort between Anthropic and Accenture plans to extend this evaluation framework to other evaluators and AI developers.