International Programme on AI Evaluation: Capabilities & Safety - Applications are open!

The first edition of the International Programme on AI Evaluation: Capabilities & Safety will run from February to May 2026. We are selecting 40 fully funded participants - graduate and PhD students, researchers, and professionals from around the world - who are ready to shape the future of AI evaluation.

We test. We measure. We make AI accountable.

Type: Fully-funded academic programme (150 hours, hybrid)

Start Date: February 2026

Location: Online + in-person capstone week in Valencia, Spain – May 2026

Priority application deadline: November 17, 2025

The Mission

The world doesn’t yet fully understand AI. We’re changing that. This programme trains the next generation of AI evaluation and safety specialists: people who will test, measure, and shape how the world trusts and governs AI.

By pioneering a new science of AI evaluation, we’re transforming scattered insights into a shared discipline and creating the standards and language that governments, labs, and societies can rely on.

What You’ll Do

  • Join 40 selected participants from around the world for a 150-hour hybrid course. 

  • Learn directly from experts at Cambridge, Stanford, Microsoft Research Asia, OpenAI, the EU AI Office, DeepMind, Apollo Research, and others.

  • Gain the foundational skills to work at AI Safety Institutes, AI companies, government evaluation teams, or research labs.

What You’ll Learn

  • Large Language Models & Architectures

  • Metrics, Benchmarks & Experimental Design

  • Red-Teaming & Adversarial Testing

  • Construct-based & Mechanistic Evaluation

  • Governance, Policy & AI Regulation
    (Full syllabus at ai-evaluation.org/programme )

Ideal Candidates

  • Graduate or PhD students, early-career researchers, or professionals, with expertise in AI and machine learning, from academic training or professional experience.

  • Curious about how we can scientifically evaluate AI capabilities and safety.

  • Excited by rigorous research, interdisciplinary thinking, and real-world impact.

Open to candidates worldwide, as we aim to build a truly global cohort representing diverse perspectives, regions, and disciplines — essential for evaluating AI systems that impact everyone.

What’s Included

  • Full scholarship funded by Open Philanthropy (tuition + travel and accommodation for the in-person week).

  • Access to an international faculty network and post-programme career pathways.

  • 15 ECTS Expert Diploma by ValgrAI - the Valencian Graduate School and Research Network of Artificial Intelligence

Selection Process

  1. Application Form – Submit your CV and motivation statements.

  2. Shortlisting – Applications will be reviewed using transparent and equitable criteria.

  3. Interviews – Shortlisted candidates will be invited for short online interviews.

  4. Final Selection – Offers extended to 40 participants, with all costs covered by Open Philanthropy.

Important Dates

  • Applications Open: October 6, 2025

  • Priority Deadline: November 17, 2025

    • Applications submitted by this date will be reviewed first.

    • Strong candidates will receive early offers before the final deadline.

    • Applications received after the priority deadline will be reviewed. However, admission is highly competitive, and spaces are limited.

  • Final Deadline: December 8, 2025

  • Final Decisions Released: January 12, 2026

  • Deadline to Accept Offers: January 19, 2026

  • Programme Starts: February 2nd, 2026 (online)

  • In-person Capstone Week in Valencia: May 25 to 29th, 2026

Shape the future of AI evaluation.

Applications are open now!

This is more than a programme. It’s a community. It’s the beginning of a discipline.
If you’re ready to do the work and make AI safer for everyone: Apply.

Abstract collage featuring a black and white moon, digital glitch effects, text snippets, and geometric shapes on a white background.

 FAQs