The specter of AI-enabled bioweapons is a wake-up call for biotech

In September 2026, leaders of major AI companies, including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, publicly warned about the serious risks posed by AI technology. Amodei argued in a September 2026 post that AI development carries serious risks and should be slowed down, while Altman agreed on social media, stating that progress should be paced. This followed the announcement by AI researcher Jacob Coxon, who left Anthropic, claiming that neither Anthropic nor OpenAI was acting responsibly. Coxon stated on social media that people building AI genuinely believe it could cause human extinction by the end of the decade. Another Anthropic employee, Evan Hubinger, publicly agreed, estimating a greater than 10% chance of human extinction from AI within the next decade.

One of the primary concerns raised by AI leaders is the potential for AI to assist in designing, creating, and releasing bioweapons. These could include highly lethal viruses targeting specific genetic groups, fungi that devastate crops causing food insecurity, or tasteless, odorless toxins introduced into water supplies. AI tools, such as molecule generators, can rapidly produce dangerous substances. For example, in 2022, researchers at Collaborations Pharmaceuticals used an AI tool to generate 40,000 potential chemical warfare agents in less than six hours. Some of these molecules were more toxic than known nerve agents, prompting the researchers to call it a wake-up call for the AI and biotech communities.

Advances in biotechnology have made gene editing and synthetic biology tools more accessible, enabling even non-experts to conduct experiments. The DIY biology movement has allowed many people to set up home labs. AI bots, trained on vast scientific knowledge, can provide instructions and video training on conducting experiments, further lowering barriers to misuse. Dunja Sabra, a biosecurity researcher at the University of Hamburg, notes that while safeguards exist, such as DNA synthesis companies screening suspicious requests, these protections are not foolproof. Sabra emphasizes that someone determined could eventually succeed in creating a bioweapon.

Several safeguards are in place to prevent misuse of biotech and AI tools. DNA synthesis companies typically screen orders for suspicious requests, and responsible researchers use red-teaming (independent analysis to identify misuse risks) and blue-teaming (developing mitigations). AI companies also modify their tools to block requests for harmful scientific information. However, Anthropic’s recent report acknowledged attempts to use its AI models to explore enhancing the chikungunya virus’s transmissibility, creating a more dangerous form of bird flu, and building an atlas of venom toxin peptides. David Magnus, a Stanford professor, highlights the ongoing challenge: AI is adept at finding ways around restrictions, necessitating continuous improvement of surveillance and screening tools.

Not all scientists agree on the level of risk posed by AI-enabled bioweapons. Some biologists at Imperial College London argue that AI tools are not yet advanced enough to fully develop bioweapons and that creating new pathogens requires time-consuming human effort. Wendy Barclay, a professor of infectious disease at Imperial, notes that the greatest pandemic risk currently comes from naturally circulating pathogens like H5N1 bird flu, which has already infected millions of birds and US dairy cattle. In contrast, Sabra advocates preparing for risks five to ten years ahead by strengthening healthcare systems, stockpiling medicines, and developing antidotes to known toxins.