Annie Jacobsen's New Book Argues Bioweapons Are Already Catastrophic—No AI Required

Annie Jacobsen's new book on biological warfare arrives at a moment when policymakers and researchers are fixated on artificial intelligence as the next great existential threat. Yet Jacobsen barely mentions AI in its pages. Her argument, as reported by Wired, is that the tools to create extinction-level biological events already exist without machine learning accelerating them.

That framing runs against the current grain. Much of the recent public debate about bioweapons has centered on whether large language models could help bad actors design dangerous pathogens, a concern that has driven new screening requirements at AI labs and prompted congressional hearings. Jacobsen's book suggests that hand-wringing over chatbots may be a distraction from capabilities that have been available in laboratories for years.

The author is known for deeply reported books on secretive national security topics, including nuclear weapons command and control and the government's continuity plans. Her method tends toward exhaustive interviews with former officials and scientists, building narratives from primary sources rather than policy abstractions. That approach gives her claims about existing biological capabilities a grounding that pure threat modeling often lacks.

Jacobsen's core point is that nature itself can be redirected with off-the-shelf techniques. Gene editing, synthetic biology, and pathogen research have advanced to the point where small teams, or even individuals with the right training, could plausibly cause harm at a scale once reserved for state weapons programs. The book does not treat this as a hypothetical future problem but as a present one.

The implications for the AI safety community are uncomfortable. If biological risk is already severe without algorithmic assistance, then regulating model weights or output filters addresses only a fraction of the danger. Biosecurity experts have argued for years that laboratory oversight, DNA synthesis screening, and international monitoring matter more than speculative AI scenarios, and Jacobsen's book gives that position a popular audience.

What happens next depends on whether Washington treats the two threats as linked or separate. Lawmakers have folded bioweapons concerns into AI executive orders and export controls, but a growing chorus of scientists says that approach misallocates attention and funding. Jacobsen's book may push the conversation back toward the biological fundamentals that no software update can fix.