The Case for ESM3 as a General-Purpose AI Model with Systemic Risk Under the EU AI Act

2026-05-02

AI Summary

This study addresses the regulatory ambiguity in the EU Artificial Intelligence Act regarding whether cutting-edge biological foundation models, such as ESM3, constitute systemic risks, leaving their dual-use potential inadequately constrained. Through a systematic analysis combining legal text interpretation, model capability assessment, and mapping of biological risk pathways, the work evaluates ESM3 against the Act’s classification criteria for general-purpose AI models. The findings reveal that ESM3 currently falls outside substantive regulatory oversight, despite exhibiting characteristics warranting systemic risk obligations. The paper argues for its inclusion under heightened regulatory duties and proposes targeted legislative amendments to integrate high-risk biological AI models into the existing legal framework, thereby strengthening governance of emerging bio-AI technologies.

Abstract

Due to ambiguity in the wording of the EU AI Act, we examine the question of to what extent frontier biological foundation models such as ESM3 are subject to obligations for general-purpose AI models with systemic risk under the EU AI Act. In this paper, we map ESM3 to the biorisk chain, and conclude that it would be desirable if the providers of ESM3 and similar biological models were subject to these obligations, which would require them to assess and mitigate dual-use risks from their models. We then perform an analysis, comparing the attributes of ESM3 to the classification criteria in the AI Act and the supporting material. We conclude that at this time, ESM3 does not appear to be meaningfully regulated by the Act. We then propose remedies to correct the situation.

Problem

• EU AI Act
• general-purpose AI
• systemic risk
• biological foundation models
• dual-use risk

Innovation

• ESM3
• EU AI Act
• general-purpose AI
• systemic risk
• dual-use risk