San Diego Alignment Workshop Convenes 300+ Researchers Ahead of NeurIPS 2025

Summary

On December 1-2, 2025, over 300 AI safety researchers gathered in San Diego for the Alignment Workshop, held immediately before the NeurIPS 2025 conference (December 2-7). The workshop focused on aligning increasingly capable AI systems with human values, with a keynote by Yoshua Bengio presenting his "Scientist AI" research program — a framework for developing non-agentic, trustworthy AI that disentangles goals from world understanding. The workshop convened alongside the broader NeurIPS 2025