"AI-designed viral genomes" is a Perspective published in Science 393(6811) on August 6, 2026 (DOI 10.1126/science.aej8512) by Thomas V. Inglesby and Moritz S. Hanke, both of the Center for Health Security in the Department of Environmental Health and Engineering at the Johns Hopkins University School of Public Health. Science published it as commentary on King et al.'s report of the first generatively designed complete bacteriophage genomes in the same issue (Generative design of bacteriophages with genome language models). Its standfirst reads: "The generation of functional viral genomes has urgent biosafety and biosecurity implications."
Argument
The opening paragraph states the position:
Artificial intelligence (AI) has been used to generate proteins, protein complexes, and short portions of genomes. Although this is promising for life sciences applications, it also raises urgent biosafety and biosecurity questions. The ability to compose viral genomes using generative AI now exists; the governance to safely steer it does not.
Two further positions were quoted in reporting published the same day: that the findings raise "urgent biosafety and biosecurity questions," and that new viruses with the potential to cause disease "should not be pursued" (Source: theguardian.com; bbc.com).
The Perspective is the counterweight in the record to the position taken by the authors of the work it comments on. Brian Hie, the senior author of both the Evo 2 paper and the phage paper, has argued that open release accelerates research, that existing pathogens are easier to obtain and produce than AI designs, and that safety checks can be engineered into an AI tool but not into natural evolution (Source: news.stanford.edu). The Evo 2 paper's own safety discussion rests on a single training-data exclusion whose limit its authors state directly — that task-specific post-training may circumvent it (Genome modelling and design across all domains of life with Evo 2).
What this page does not record
AAAS paywalls the body of the Perspective. The title, standfirst, opening paragraph, both authors and their shared affiliation, the publication date and the article type were retrieved verbatim from the publisher's DOI landing page; the remainder was not retrievable. Whatever specific governance mechanisms Inglesby and Hanke propose — screening regimes for synthetic nucleic-acid orders, pre-publication review, model-release conditions, or an oversight body — is therefore not recorded here, and no positive recommendation should be attributed to them from this page. Confidence is held at medium for that reason. Retrieving the full text would let the page state their proposals rather than only their objection.