ESM3 Biological AI Model Found Largely Unregulated by EU AI Act
A new paper argues that ESM3, a frontier biological foundation model, is not meaningfully regulated under the EU AI Act despite its potential for dual-use risks. The authors map ESM3 to the biorisk chain and conclude that the model's providers should be subject to obligations for general-purpose AI models with systemic risk, but current classification criteria fail to capture it. They propose remedies to close the regulatory gap.
Key facts
- ESM3 is a frontier biological foundation model.
- The paper examines ESM3 under the EU AI Act.
- ESM3 is mapped to the biorisk chain.
- The authors conclude ESM3 is not meaningfully regulated.
- Providers should assess and mitigate dual-use risks.
- The paper proposes remedies to correct the situation.
- The analysis compares ESM3 to AI Act classification criteria.
- The paper is published on arXiv.
Entities
Institutions
- arXiv