Hinton, Li and Ng Back AI Openness While Dividing Over Open-Weight Risks

Geoffrey Hinton, Fei-Fei Li and Andrew Ng used a joint appearance at Ai4 2026 to argue that artificial intelligence must remain broadly accessible, even as they split over how freely trained model weights should circulate and how regulation should address the resulting risks. The discussion exposed a central policy dilemma: preventing harmful use without concentrating control of a foundational technology inside a few companies.