Security Considerations at the Intersection of Engineering Biology and Artificial Intelligence
Engineering Biology Research Consortium; Compiled And Edited By Charlie D. Johnson, Wilson Sinclair, Rebecca Mackelprang · 2023
Engineering biology and artificial intelligence (AI) ii are both characterized by their rapid growth, their potential to dramatically accelerate technological solutions to pressing challenges, and their tendency to usher in novel security concerns.It is unsurprising, then, that the emerging capabilities at the intersection of these technologies are generating both great enthusiasm and great concern.As the power of this technological convergence is still being explored and realized, it is timely to identify and consider potential security implications and how they might be addressed.By engaging in such a process now, researchers, policymakers, and members of society iii can consider carefully what acceptable levels of risk given likely benefits might-be and for whom-as new capabilities are developed.i The Engineering Biology Research Consortium brings engineering biology researchers and other stakeholders from industry and academia together with policymakers to advance engineering biology to address national and global needs.EBRC White Papers are developed with significant, substantive direction, guidance, and input from EBRC members through an interactive process.EBRC products do not necessarily reflect the direct views of all EBRC members.ii Artificial Intelligence is an umbrella term that encompasses generative AI, many types of machine learning, and deep learning.To avoid confusion and in recognition that different AI-based approaches can be used to address the same challenges, the authors use "AI" to refer collectively to these approaches.iii This paper was developed by EBRC members, who are generally based in the technical research community.Thus, this paper takes a technical approach to this topic.However, members also pointed to the present opportunity to think more broadly about societal impacts, perceptions of risk, and roles in the determination of technological Overview & Key RecommendationAs researchers and innovators work to leverage the power and opportunity of artificial intelligence to advance engineering biology to address societal challenges, stakeholders must also recognize the potential of AI-enabled capabilities to cause harm.Herein, we consider three areas-de novo biological design, closed loop-autonomous laboratory systems, and natural language Large Language Modelswhere the intersection of these technologies may pose significant security concerns.For each, we describe the technology, associated security concerns, and how those concerns might be addressed.We recommend that a regularly convened international forum or initiative be established to bring diverse stakeholders together to identify security concerns at the intersection of engineering biology and AI and consider how they might be addressed.