NTI | bio Advances Responsible AI Biodesign Tool Development and Use

The integration of artificial intelligence (AI) with the life sciences offers tremendous potential benefits to society, but advances in AI biodesign tools also pose significant risks of misuse, with the potential for global consequences.

AI biodesign tools (BDTs) — technologies that enable the engineering of biological systems — have the potential to drive progress in the development of new therapeutics and are likely to have an impact across the broader bioeconomy, including in agriculture, health, and materials science. “As these tools become more capable of designing novel biological systems—with tremendous potential benefits for basic research as well as applications in vaccine development and other key areas—we should also be cognizant of the risks that BDTs could be exploited to design dangerous pathogens,” noted Jaime Yassif, NTI | bio vice president. “Currently, very few safeguards exist to ensure that the benefits of these technologies can be realized safely and securely.”

NTI | bio’s new report, Developing Guardrails for AI Biodesign Tools, outlines several risk reduction measures that could be deployed to reduce the risks associated with biodesign tool development and release. This paper builds on NTI | bio’s previous report, The Convergence of AI and the Life Sciences, which highlighted this gap and called for research to identify promising new governance approaches to safeguard large language models, automated science technologies, and biodesign tools.

Biosecurity experts, biotechnology industry leaders, NGOs, governments, and others have worked to assess and reduce biological risks related to the misuse of AI tools. However, many of these discussions have focused on large language models, not the specialized biodesign tools that could soon become capable of enabling researchers to design novel pathogens that can spread more easily among human populations or cause more severe health damage. Developing Guardrails for AI Biodesign Tools explores a range of guardrails that could be developed to help safeguard biological design tools and offers recommendations in two focus areas:

• Built-in guardrails, which refer to technical solutions for risk reduction that can be included in the development or use of a BDT.

• Managed access paradigms, which includes ways to provide differential access to BDT software—through a range of approaches on the continuum between open-source and closed models—based on the needs of different developers and users.

Governments, the private sector, and the biosecurity community have a lot of work to do to support the development of safe and secure biodesign tools, and this new report provides recommendations for pilot projects to make progress on these important risk reduction measures. These projects include, among other approaches:

• Coupling metadata with biological designs to inform DNA synthesis providers of the origin and intent of a new design.

• Establishing and supporting a managed access platform for BDTs to provide additional resources and access to the scientific community while managing risks.

• Developing an ecosystem to support screening and additional review of new biological designs with high potential for misuse, building on the work of the DNA synthesis screening community.

The built-in guardrails and managed access approaches outlined in the report offer a number of possible guardrails to reduce risk. “Implementing these strategies though a range of pilot projects can determine the feasibility of these approaches to reduce risks while supporting innovation,” explained Sarah Carter, a lead author of the report.

NTI | bio will continue to advance technical development in this field by pursuing pilot projects and will continue to convene expert discussions to explore novel governance approaches to AI in the life sciences, including through the next meeting of the AI-Bio Global Forum in December 2024.

Read the full report here.