Top Executives Sound Alarm on AI Biosecurity Risks

AI could be used to create biological weapons. Faced with this reality, top executives in the AI and biotech industries have issued warnings about the risks. Over the weekend, Anthropic CEO Dario Amodei argued that AI poses serious dangers and that its pace of development should slow, while OpenAI CEO Sam Altman posted on X (formerly Twitter): "Like Dario, I agree that we need to slow the pace of frontier AI."

A few days earlier, AI researcher Jacob Hilton left his position at Anthropic, claiming that Anthropic and OpenAI were not acting responsibly, and posted about it on X. Anthropic employee Evan Hubinger said on X: "I seriously believe AI could cause human extinction. I personally think the probability within the next decade exceeds 10%."

One of the doomsday scenarios these figures fear is AI being involved in the design of bioweapons. MIT Technology Review (Jessica Hamzelou, September 18, 2026) reported that AI has made the design of a wide range of bioweapons easier than ever — from deadly viruses engineered to target specific populations based on their genetics, to fungi that wither crops and trigger food crises.

In 2022, researchers at Collaboration Pharmaceuticals reversed and tested an AI molecular generation model they had developed to find candidate compounds for disease treatments. Within six hours, the model produced 40,000 molecules that could serve as candidates for chemical warfare agents, some of which were designed to be more toxic than known nerve agents. At the time, the researchers wrote: "Without being alarmist, this should be a wake-up call for our colleagues in the AI drug discovery community."

David Magnus, a professor of medicine and biomedical ethics at Stanford University, called that report "extremely frightening." An expert who has been assessing the risks of misuse in medicine and biotech since the late 1990s, he said: "That was truly terrifying, and the situation since then has grown beyond control."