The Fermentation Tank Problem — Why Bioweapons Don't Need Secrecy

Annie Jacobsen's new book on bioweapons ignores artificial intelligence entirely — not because she missed the angle. Because the older threat already works.

Scientists have known since the 1970s that you do not need a weapons laboratory to turn pathogens lethal, only fermentation tanks, published microbiology textbooks, and time. In 1972, after researchers demonstrated exactly this, 143 nations signed the Biological Weapons Convention and agreed to stop.

The treaty banned offensive bioweapons research but created no enforcement mechanism. What happened next is the structural trap Jacobsen is watching repeat.

The BWC collapsed not because nations kept forbidden bioweapon programs — they certainly did. But because the distinction between defensive research and weaponizable research cannot be enforced in real time. A fermentation facility built to study pandemic flu mitigation is indistinguishable from one built to weaponize it. Gain-of-function research, conducted openly at universities and funded by governments under the label "academic," steadily improved pathogen transmissibility and lethality without crossing any verifiable line. Enforcement required inspectors inside laboratories. Nations refused entry.

Today the baseline risk is not that someone will build a new pathogen in secret. It's that someone will build it in a published paper, under grant funding, with proper citations. Call it science. Gain-of-function research continues. Synthetic biology costs less each year. Jacobsen's omission of AI suggests a pragmatic choice. She chooses to worry about biology that is already weaponizable now, in laboratories that exist, with knowledge already published and spreading. The reader who wants to act usefully needs to see what is already true before imagining what might become true.