Bioweapons, AI, and the end of the world, with Annie Jacobsen
Bremmer warns that weaker global cooperation could undermine responses to catastrophic threats, while AI may put specialist knowledge within reach of bad actors. Jacobsen argues that biological weapons have lower barriers to entry than nuclear weapons and that AI could accelerate biological-weapon research, though pathogen sensors are not yet a ready defense. The discussion also touches on AI safeguards, the incentives of AI companies, and the history and regulation of biological weapons.
Preview · 11 more points in the app
Read the Full SummaryKey Takeaways
Jacobsen says biological weapons have a lower barrier to entry than nuclear weapons, and that an airborne pathogen can use the human lung as its delivery system. She also notes that ubiquitous pathogens are harder to control than smallpox, whose attempted procurement or possession can carry a $2 million fine and 25 years in federal prison.
Jacobsen argues that AI could lower the expertise needed for biological-weapon research by rapidly scanning documentation and performing computational trial and error; she describes AI as providing a blueprint rather than creating a weapon itself. Bremmer says Anthropic blocked some users seeking potentially useful biological-weapon information, though the company did not know their intent.
Jacobsen says ChatGPT initially refused her biological-weapon request, but that a new account and differently phrased questions were not blocked. Her account illustrates a way she bypassed the refusal, underscoring the limits of that safeguard as she described it.
GEO-POLITICS
“Or another example from Iran. Anthropic says that the Iranians used Claude to help track US Navy ships in The Middle East.”—
“But imagine my shock literally when I was looking through defense department threat spectrums and learned that not only is biological war extremely close on that threat spectrum to nuclear…”—
“The barrier to entry is extremely low. You do not need any of the material that is so closely guarded by nuclear command and controls.”—
TECHNOLOGY
“But the world today is increasingly fragmented…”—
“Recently, Anthropic, one of the leading AI companies in The United States revealed that it had blocked several users of its popular AI platform, Claude, from using its models to…”—
“That's good old r and d. And this would go on, you know, trial and error for decades…”—

Catastrophic Risk And AI
Bremmer warns that fragmented global cooperation and AI-enabled access to specialist knowledge could make existing catastrophic threats harder to contain.
Biological Weapons And AI
Jacobsen argues that biological weapons have lower barriers to entry than nuclear weapons, and that AI can reduce the expertise required to plan them.
Bioweapon Materials And Methods
Jacobsen recounts a historical biological-weapon plot and explains why accessible pathogens and equipment make the threat more plausible today.
AI Company Incentives
Bremmer and Jacobsen discuss government oversight, companies’ claims about AI risks, financial incentives, and the value of experienced experts.
Biological And Existential Risk
Jacobsen connects biological and nuclear dangers to the pace of technological change and argues that experienced experts should help guide future safeguards.
Related Episodes



Important Disclaimers
Not Investment Advice: For education only — not financial, tax or legal advice.
AI-Generated: Made by AI from source transcripts and may contain errors.
No Guarantees: Past performance does not predict future results; investing carries risk.
Our Commentary · Request Removal