Artificial Intelligence Enabled Viral Weapons of Mass Destruction, A Critical Perspective on The Current Landscape of Nucleic Screening, AI Agents, and Implementing the National Artificial Intelligence Action Plan

Abstract

The current trajectory of synthetic biology and AI development, while promising significant societal benefits, is simultaneously creating critical biosecurity vulnerabilities that our current tools and governance structures are ill-equipped to handle. This paper highlights gaps surrounding viral weapons of mass destruction (VWMDs), and provides an overview of key policy, technical, and regulatory barriers.

Currently, there is a lack of enforced regulations to effectively screen customers purchasing biosynthetic products, and a vulnerability within existing screening processes to adversarial exploitation. At the same time, global governance and policy frameworks have poor oversight over private companies and a lack of incentives to promote biosafety. We suggest a strategic, top-down approach that focuses on developing effective screening tools and pragmatic customer screening policies. We put forward key considerations and opportunities for the EO’s upcoming AI Action Plan to address current gaps.