Science Weekly Science Weekly

‘These risks are real’: the threat of AI bioweapons

Oct 20, 2026 · 15m

Summary

Science Weekly explores the emerging threat of AI-enhanced bioweapons with Dr. Tom Inglesby, director of the Johns Hopkins Center for Health Security. The discussion covers how AI models can design novel pathogens and lower the skill barrier for bad actors, citing recent examples of state-sponsored attempts to modify viruses. Inglesby outlines critical safeguards, including access controls for sensitive data, mandatory government evaluations of powerful models, and screening of nucleic acid orders to prevent physical synthesis.

Topics discussed

Intro: AI risks and bioweapons threat Defining bioweapons and historical context When AI-bioweapon concerns emerged Nation-state risks and novel virus design Motivations and reported state attempts Uplift risk for individuals and terrorists Accidental creation of dangerous pathogens Proposed safeguards and access controls Dr. Inglesby's assessment of the threat Outro and sponsor segment

More from Science Weekly

Could measles and midterms bring down Maha?
Nov 19, 2026
Nasa’s Voyager 1 probe in sight of an incredible milestone
Nov 17, 2026
Is BMI working? Cop31 latest, the science of great conversation
Nov 12, 2026
Why we get Sad and what helps
Nov 10, 2026
‘Ferritin face’: can you really diagnose low iron from a selfie?
Nov 5, 2026
Listen ad-free on Castria