Nuclear weapons have long represented the ultimate doomsday threat. But investigative journalist and author Annie Jacobsen says biological weapons may pose an even greater risk, in part because they require far fewer resources and specialized capabilities to develop.

On GZERO World, Jacobsen joins Ian Bremmer to explain how artificial intelligence could lower those barriers even further, giving users access to expertise that once required years of specialized training. They discuss how a bad actor could move from an AI-generated blueprint toward a real-world threat and where effective safeguards might actually work.

They also examine gaps in current AI guardrails, the struggle between government regulation and rapidly advancing private technology, and the broader problem of human innovation moving faster than our ability to manage its risks.