Stanford Medicine

AI, Biology, and Biosecurity in the Age of Acceleration | Stanford's RAISE Health Symposium 2026

At Stanford’s RAISE Health Symposium 2026, speakers warned that artificial intelligence, now a general‑purpose technology, is also a dual‑use tool that can transform both medicine and biological weapons. The presenter contrasted AI’s rapid, worldwide diffusion with the 1975 Asilomar meeting on recombinant DNA, noting that the old consensus‑building model cannot contain today’s AI. While AI‑driven models have already identified promising new antibiotics from archaea, the same systems can design potent nerve agents, synthesize pandemic pathogens, and poison scientific literature. Examples cited included a toxicity‑trained model that suggested compounds surpassing VX, an AI tutorial that taught non‑scientists how to acquire a viral agent, and Stanford researcher Brian He’s genome‑language model that built functional bacteriophages—demonstrating both therapeutic promise and the ease with which malicious actors could replicate the work. The speaker argued that without a robust, adaptive governance framework—combining top‑down regulation, bottom‑up oversight, access controls, and genuine public deliberation—society faces a widening cultural lag. Slowing AI deployment or accelerating policy responses is essential to prevent catastrophic biosecurity outcomes while preserving legitimate medical advances.