Anthropic Shares Framework to Track AI Development Pace

The framework measures automated AI R&D progress, AI agent oversight quality, and internal compute allocation.

Anthropic published an initial snapshot of the three metrics based on internal company data.
The methodology is designed for other frontier AI developers to adopt and for third parties to verify.
The metrics aim to standardize transparency around how AI systems contribute to building subsequent AI models.

Policymakers and researchers get a proposed standard framework to monitor automated AI development and resource distribution across frontier labs.

Sources