Anthropic published three metrics for transparency in AI development

Anthropic today published an internal snapshot and methodology proposing three metrics to track AI development: how much research and development is performed by AI, how effective oversight of AI agents is, and how computational capacity is distributed. The aim is to reduce the gap between frontier laboratories’ knowledge and public understanding so that society can make informed decisions about next steps.