Fazl Barez

Projects

Research Direction: Automated Interpretability

• AI Agents for Interpretability

• Training Models to explain their own computations

• Empowering Users through Explainability

What I'm looking for in a Mentee

• Strong technical background in AI and ML

• Good programming skills

• previous experience in ai safety/interpretability would be a plus

• Independent and motivated

• Quick learner

Bio

Fazl Barez is a Senior Research Fellow at the University of Oxford dedicated to building AI that strengthens human agency. His work spans interpretability, safety, unlearning, and governance, developing both technical systems and institutional safeguards that allow people to understand, shape, and direct powerful AI models. His research is motivated by a simple principle: AI should amplify human judgment and autonomy — not replace it, obscure it, or centralise it in the hands of a few.