Open Philanthropy is launching an RFP for work on AI capability evaluations. We're looking to fund three types of work:

  1. Global Catastrophic Risk (GCR)-relevant capability benchmarks for AI agents
  2. Research to improve our understanding of how capabilities develop and scale
  3. Solutions for enabling meaningful third-party evaluation while maintaining security

Key details:

Even if you're not planning to apply, the RFP contains several open research questions that we think are important for the field. Consider applying if you have relevant expertise, and please share with others who might be interested.

New Comment
More from cb
Curated and popular this week