AI is reviving fears around bioterrorism. What's the real risk?

Abstract

As frontier language models become more capable at chemical and biological queries, concerns about AI-enabled bioterrorism have re-entered mainstream security discourse. Drawing on the International AI Safety Report and recent demonstrations of chatbots advising users on how to plan attacks using novel bacteria, viruses, and toxins, the article assesses the real-world plausibility that extremists or other malicious non-state actors could weaponize such capabilities to advance ideological causes. It argues that great-power competition over frontier AI is distracting policymakers from the more immediate threat posed by terrorist groups, and calls for renewed attention to biosecurity guardrails on general-purpose AI systems. – AI-generated abstract.