Thao Amelia Pham

I work on problems of cooperative AI, specifically, techniques and foundational methods to ensure advanced AI agents cooperate and navigate the real-world environment safely. Some research directions I'm working on include:

• Verification of cooperation in large-scale multi-agent AI systems
  • Can game-theoretic and simulation-based methods to detect collusion and emergent misalignment?
• Transparency and oversight mechanisms in adversarial multi-agent environments
  • Whether they improve coordination or introduce covert communication risks?
• Reinforcement-learning approaches for inducing robust prosocial behaviour

I'm originally from Vietnam. I'm finishing up my B.A. in Mathematics and Computer Science in the United States. Connect with me via email.