Building the Field of AI Evaluation — Inside the Programme, Year 1

The International Programme on AI Evaluation: Capabilities & Safety is the first global academic programme dedicated to evaluating AI systems. Its first cohort brought together 40 participants — selected from over 670 applicants — across machine learning, psychometrics, law, policy, and beyond. From February to June 2026, they worked through ten online modules and a capstone project, culminating in a week together in Valencia: presenting their work, meeting their mentors and peers in person, and shaping what rigorous AI evaluation can look like. This is a look at that journey, and the people building the field. Run in partnership with ValgrAI, Berkeley Existential Risk Initiative and Coefficient Giving. ai-evaluation.org