Shanghai, June 16 2026

IBBIS and Concordia AI held a closed-door workshop on AI and biosecurity governance during the World AI Conference (WAIC) in Shanghai. The session brought together researchers, policymakers, and industry experts to work through three questions about how AI is reshaping biosecurity risk, which mitigations hold up in practice and how international governance of AI-bio convergence might take shape.

Executive Director Piers Millett spoke on risk mitigation. He made the case that AI model-level safeguards and screening at the point of DNA synthesis are strongest when they reinforce each other, since neither closes the gap on its own.

The argument builds on research by IBBIS and Microsoft, published in Science, which found that AI protein design tools can generate variants of proteins of concern that slip past the screening systems used by DNA synthesis providers. Screening at the point of synthesis catches dangerous sequences before they are ever made. Safeguards inside AI models limit what those tools will design in the first place. Each covers a gap the other leaves open.

China is one of the world’s largest DNA synthesis markets and a major site of AI development. A screening system that works only in some regions leaves an opening everywhere else, which makes engagement in China central to any approach meant to hold globally. The workshop continues IBBIS’s engagement with stakeholders across China and the wider region.

Learn more about IBBIS’s work on DNA synthesis screening here.