Rapid AI and biotechnology development brings promising health benefits but creates unprecedented biosecurity risks, which current global treaties and data systems cannot sufficiently address. The RAND–NTI workshop at the 2025 AI Action Summit convened stakeholders from multiple sectors to examine challenges and responses. Key discussions centered on gaps in reliable data, policy obstacles, and the need for coordinated measures to track, contain, and manage AI-driven biosecurity threats.

Biosecurity Governance Across Uncertain Artificial Intelligence Futures

Perspectives from a Side Event on Biosecurity and Frontier AI at the 2025 AI Action Summit

Aurelia Attal-Juncqua, Casey Mahoney, Sana Zakaria, Paul Khullar, Patricia Paskov, Ella Guest, Michael Simpson

Expert InsightsPublished Sep 24, 2025

Advances in artificial intelligence (AI) offer tremendous potential benefits for science and medicine, but also introduce new risks which must be thoughtfully managed to safeguard global health and security. RAND partnered with the Nuclear Threat Initiative (NTI) to co-convene a workshop on the sidelines of the AI Action Summit in Paris on 11 February 2025 in order to assess the perspectives of subject-matter experts on emerging challenges at the nexus of AI and biology, and to identify solutions. The event, ‘AI Safety Institutes and AIxBio Governance: A Discussion with Biosecurity Experts and AI Model Developers’, brought together policymakers, technical experts, and public- and private-sector stakeholders from advanced industrial economies. This report summarises key insights from the workshop and has been prepared by RAND’s Technology and Security Policy Center and Frontiers of Technology Hub.

Recommendations

Workshop participants identified practical steps that stakeholders can take to reduce risks, close knowledge gaps, and build more resilient systems. Key takeaways included:

  • Focus on measurable risks. Participants recommended shifting from broad scenarios to specific, measurable indicators, using ‘if–then’ hazard-threshold frameworks to enable timely mitigation even when full risk quantification is impractical.
  • Invest in risk assessment and testing. Participants called for the development of collaborative methods and safe proxy environments to evaluate misuse-relevant AI pathways, and test both general and specialised AI tools against misuse scenarios affecting healthcare delivery or public infrastructure.
  • Strengthen policy adaptability. Experts emphasised the need for governance systems that evolve with technology, using feedback loops among researchers, policymakers, and civil society to update safeguards as risks change.
  • Promote cross-sector coordination and information sharing. Participants agreed that improving communication and privacy-preserving information sharing between AI and biosecurity communities is key to effective early warning and response.
  • Support safety-driven innovation. The group discussed the co-development of technical solutions like access controls, secure compute, and auditing to reinforce biosafety norms.
  • Keep collaboration going. Participants found the RAND-NTI partnership valuable in highlighting diverse perspectives and urged continued efforts to build bridges across sectors and sustain global attention concerning AI-related biosecurity risks.

Topics

Document Details

Citation

Chicago Manual of Style

Attal-Juncqua, Aurelia, Casey Mahoney, Sana Zakaria, Paul Khullar, Patricia Paskov, Ella Guest, and Michael Simpson, Biosecurity Governance Across Uncertain Artificial Intelligence Futures: Perspectives from a Side Event on Biosecurity and Frontier AI at the 2025 AI Action Summit. Santa Monica, CA: RAND Corporation, 2025. https://www.rand.org/pubs/conf_proceedings/CFA4186-1.html.
BibTeX RIS

Research conducted by

This work was conducted using income from operations and gifts from philanthropic supporters. A complete list of donors and funders is available at www.rand.org/TASP.

This publication is part of the RAND conference proceeding series. Conference proceedings present a collection of papers delivered at a conference or a summary of the conference.

This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.