Harvard fellow 'deeply concerned' over efforts to build bioweapons after Anthropic report
Dr. Ashish Jha, a Harvard Senior Fellow, expressed deep concern over the increasing use of AI, particularly open-source models, to develop biological weapons. He noted that while major AI companies like Anthropic have implemented guardrails, many open-source models lack such restrictions and have been downloaded millions of times globally. Jha highlighted a case where an individual, after being blocked by Claude, turned to an open-source model to pursue bioweapon development, emphasizing the urgent need for strengthened biodefense strategies.
The increasing accessibility of AI tools for developing bioweapons poses a significant threat to global security, potentially empowering non-state actors and terrorist groups with the ability to create destructive biological agents.