Biosecurity

Biosecurity is the discipline, primarily of policy and practice, intended to prevent biological agents, biological knowledge, and biological tools from being used to cause harm — whether through deliberate theft, insider misuse, dual-use research outcomes, or hostile-state or non-state actors. The term emerged in its modern form following the 1996 adoption of the Biological Weapons Convention's enforcement protocols and the subsequent framing of biology as a national-security issue, alongside related concerns about laboratory biosafety. Biosecurity now spans multiple communities — laboratory, policy, intelligence, public health — and has been steadily integrated into international legal frameworks.

The mechanism of biosecurity in laboratory and research settings centers on three practices: screening of orders for synthetic DNA so that dangerous sequences are not synthesized without review; personnel reliability programs that assess researchers with access to sensitive biological materials; and secure handling of biological select agents and toxins, particularly those with potential dual-use applications such as the major human and animal pathogens. At the national and international level, biosecurity also covers export controls on biological equipment and expertise, the implementation of the Biological Weapons Convention, and intelligence work on emerging biological threats. The practice of screening orders has been a particular focus of synthetic-biology policy since the early 2010s, as DNA synthesis became commercially accessible and as new synthesis techniques produced longer and longer constructs.

The tradeoffs in biosecurity are political and practical. Strong screening and personnel programs reduce the rate at which biological tools can spread, at the cost of slowing legitimate research and creating friction at the lab-industry interface. Biosecurity rules are also necessarily incomplete: they tend to focus on well-characterized pathogens and known sequences, while AI-designed novel constructs may not match any existing risk classification. International harmonization is uneven, with significant variation in screening rules across jurisdictions. The rise of generative biology and autonomous wet labs in the mid-2020s has made these gap areas more acute: the synthesis step is being automated, the design is being delegated to AI systems, and the human actor with the directest control over what gets made is increasingly distant from the wet lab itself.

Open questions include how to extend existing biosecurity frameworks to cover AI-generated biological designs that may include novel functions not captured by current sequence-based screening; how to govern research that is geographically distributed across many legal regimes; and how to decide which functions in engineered organisms count as dual-use. There is no consensus on whether the appropriate response is to extend screening criteria to AI outputs, to require pre-publication review for AI-designed biology, to monitor compute and DNA-synthesis rather than the biology itself, or to leave the question to professional norms. The connection to biosafety is increasingly discussed in joint frameworks, and to AI governance in general is starting to be made in policy proposals at national and international levels.