To keep AIs safe, Anthropic is offering grants to third parties to develop “benchmarks” that will assess the effectiveness and potential threats from new AI models.
As AI Rises, Startup Anthropic Wants Other People to Invent Tools to Measure if it’s Safe or Threatening.
