AI models design viruses not found in nature for first time

Lead

US researchers have announced a scientific first: the successful synthesis of brand-new viruses not found in nature, based entirely on designs generated by artificial intelligence. The viruses, known as bacteriophages, are fully functional and can replicate in the laboratory, marking the first time whole genomes have been designed by AI. While the breakthrough offers potential benefits for treating disease, it has also prompted warnings from experts about urgent safety and security implications.

Coverage Comparison

Reporting on the story comes from two major outlets: ABC Australia and BBC — World. Both emphasise the novelty of the achievement, though their framing differs slightly in focus.

ABC Australia highlights the safety and security concerns, noting that the ability to compose viral genomes using generative AI now exists while governance to safely steer it does not. The outlet describes the development as a cautionary milestone, pointing out a "huge disconnect" between technological capability and oversight.

BBC — World takes a somewhat more neutral tone, framing the news as a breakthrough that could unlock a new era for treating disease. The outlet quotes a researcher who called it "a very significant turning point" in science, while also carrying expert warnings about "urgent" safety and security concerns. BBC — World also explains that the technology works similarly to large language models like ChatGPT, which predict sequences of text.

Despite these slight differences in emphasis, both sources agree on the core facts: that AI has been used to design viable viruses, that the viruses pose no threat to people, and that the achievement carries both promise and risk.

Key Claims

AI-Designed Viruses Are Fully Functional — According to BBC — World, artificial intelligence has been used to design brand new viruses that are fully functional and can replicate in the laboratory. This is the first time whole genomes have been successfully designed by AI.

16 Novel Viruses Created — The resulting 16 novel viruses were created to infect bacteria and pose no threat to people. The researchers synthesised the most promising designs and tested them against E. coli bacteria.

A Significant Turning Point — The breakthrough has been labelled a "very significant turning point" in science, according to BBC — World, potentially unlocking a new era for treating disease. The outlet quotes Brian Hie, assistant professor at Stanford University, who said, "This is a next step in the complexity that's designable by generative AI, this is the first time generative AI has been used to design a complete genome."

Urgent Safety and Security Concerns — Experts have warned that AI-designed viruses raise "urgent" safety and security concerns, as reported by both ABC Australia and BBC — World. The concern is that while the technology advances rapidly, governance lags behind.

How the Technology Works — The technology operates similarly to large language models like ChatGPT, which predict sequences of text. In this case, the AI models predict genetic code rather than words, having been trained on genomes from viruses, bacteria, plants, and animals.

The Role of Evo 1 and Evo 2 — ABC Australia reports that researchers used two "genome language models" named Evo 1 and Evo 2 to generate complete genomes for a viable bacteriophage. The models were given an existing bacteriophage, ΦX174, as an example to mimic.

Scale of Design Attempts — The models generated about 700,000 potential designs, of which the researchers picked 285 that looked most promising. BBC — World, however, notes that the Stanford researchers picked 302 of the most promising designs for synthesis, yielding 16 effective viruses. This discrepancy in numbers remains unexplained.

Demonstrated Viability — In the lab, the researchers synthesised DNA molecules based on the designs and inserted them into E. coli bacteria. Within a short period, 16 of the Petri dishes showed clear spots as the viruses attacked and replicated inside the bacteria, proving their viability. Some of the viable viruses proved more effective at attacking E. coli than the original ΦX174.

Perspectives

Researchers and Proponents — Brian Hie, assistant professor at Stanford University, described the work as new territory for generative AI, emphasising that this is a step forward in the complexity of what can be designed. Samuel King, a PhD student in the lab, reportedly said they realised the phage designs were effective, according to BBC — World's report.

Safety and Security Experts — While specific experts are not named in the provided material, both sources relay warnings from unnamed experts about "urgent" safety and security concerns. These concerns centre on the potential misuse of AI-designed viruses, given the rapid advancement of technology versus the slow pace of governance.