The AI So-Go-Chi Center held a research seminar on July 16, 2026, featuring Dr. Emeric Alexander Neufeld, who is currently a JSPS Postdoctoral Fellow (Summer Program) at the AI So-Go-Chi Center, hosted by Associate Professor Teeradaj Racharak.

In his talk, “From Safety to Compliance: Teaching AI Legal and Ethical Behaviour,” Dr. Neufeld discussed an important aspect of trustworthy AI: enabling autonomous agents to comply with legal, ethical, and social norms. He introduced recent advances in integrating normative reasoning with reinforcement learning, demonstrating how techniques originally developed for AI safety can be extended to support compliance with potentially conflicting norms and priorities. The seminar also highlighted the use of formal safety specifications, multi-objective reinforcement learning, and deep reinforcement learning to develop autonomous agents capable of making decisions in complex stochastic environments.

The seminar concluded with an active discussion on trustworthy AI, normative reasoning, and future directions for integrating formal methods and machine learning in autonomous systems.