Dual-Use AI

AI systems or capabilities with legitimate civilian applications that can also be repurposed for harmful ends. Examples include AI-generated synthetic media (deepfakes), AI-assisted cybersecurity tools, and AI for drug discovery that could enable bioweapons. Dual-use risks require governance approaches that balance enabling beneficial uses with preventing harms.