Bans on Open-Weight Models Fail to Deter Malicious Use

Boaz Barak, a theoretical computer scientist, Harvard professor, and OpenAI alignment researcher, argued that prohibiting open-weight models will not eliminate them or stop misuse. He noted that individuals intent on harmful applications would likely disregard any regulations. AI YouTuber Matthew Berman agreed that such bans would not address risks from non-compliant actors and suggested focusing enforcement on large-scale model distillation operations instead. The discussion highlights tensions between AI safety proposals and the practical limits of restricting publicly available model weights.