not be dismissed, says Luca Righetti, another author of the study, who conducted the work while at METR, an AI-safety group. And technical progress continues. Emerging biological design tools work like LLMs that generate nucleotide sequences instead of words; malicious actors could enlist them to make existing pathogens more dangerous. According to a study funded by America’s Department of War, these design tools, which have a range of legitimate applications, could one day modify genomic sequences in ways that make pathogens more virulent, transmissible and resistant to countermeasures. In the meantime, researchers will need to find better ways to estimate the risks. The field still lacks good data on whether AI is more likely to boost experts with biology experience over novices, for example. Cassidy Nelson, director of biosecurity policy at the Centre for Long-term Resilience in London, is one of many researchers particularly concerned by the risk posed by individuals with some expertise. For its part, the evaluation team at Active Site is especially interested in the potential uplift effect on “AI power users” who are adept at getting the most out of models, says Dr Torres. Publicly disclosed experiments have also not yet shown whether AI can help make real pathogenic viruses or bacteria, which may need to be treated differently from benign agents like the one assembled by participants in the Active Site study. Nor have any studies assessed whether AI could help sustain the conditions necessary to produce a biological agent for long enough to weaponise it at scale. Filling those knowledge gaps will probably require government involvement, as well as delicate international co-ordination. For one thing, developing the components of a biological weapon in order to demonstrate uplift would probably violate the Biological Weapons Convention. Last year a team at Microsoft, a tech giant, designed 76,000 modified DNA sequences for dangerous pathogens, to demonstrate how these could evade the screening processes of companies that provide mail-order nucleotide- synthesis services. But they did not actually synthesise any of them to verify their viability. Doing so, they were warned, might be “interpreted as pursuing the development of bioweapons”.

Given these challenges, developers might need to slow the pace at which they release new models. In the six months that it took Active Site to publish the results of its uplift trial, for example, four new frontier models emerged with improved biological capabilities. Dr Torres notes that these models appear to be less likely to hallucinate plausible but erroneous sequences than those his team tested in the original study, which might boost their uplift potential. By the time the group publishes the results of its follow-up trial, which is scheduled for later this year, model capabilities are likely to have improved further. There is precedent for such caution. Last month Anthropic announced that it was limiting access to Mythos, its world-leading cyber-security model, until the risks it poses could be resolved. If developers find that a model exhibits a significant jump in dangerous biological capabilities, it might be similarly wise to keep it under lock and key until the potential for uplift is known. With stakes as high as these, a little patience could go a long way. ■ Curious about the world? To enjoy our mind-expanding science coverage, sign up to Simply Science, our weekly subscriber-only newsletter. This article was downloaded by zlibrary from https://www.economist.com//science-and-technology/2026/05/05/how-ai-tools-could- enable-bioterrorism

Science & technology | Troubled waters How worried should you be about hantavirus? An outbreak on a cruise ship has authorities concerned May 7th 2026 ON APRIL 1ST MV Hondius, a cruise ship carrying around 150 passengers and crew, set sail from Argentina towards the island nation of Cape Verde. By early May an outbreak of hantavirus was reported on board, sending international health authorities scrambling to contain further spread and treat those taken ill. As of May 6th three cases among those on board had been confirmed and five more were suspected. Of those eight people, three have died. Although information about the outbreak is still patchy, the World Health Organisation (WHO) and other health bodies say the risk of hantavirus infections globally remains low.

Human infections with hantavirus are rare. Such cases are usually caused by the inhalation of airborne particles from the droppings or urine of rodents such as mice and rats, in which the virus is endemic. Further spread from infected humans to others is rarer still, though not unheard of. Very close contact—such as bed-sharing, sex and interactions between health-care workers and patients—has historically been a prerequisite. Although no human-to-human transmission on MV Hondius has yet been confirmed, the cramped cabins and common areas of a cruise ship are an ideal environment for it to occur. The fact that the wife of the first passenger who died was also infected is a worrying sign. There are many strains of hantavirus, each harboured by different animal species. At least three of the infected individuals from MV Hondius have been diagnosed with the Andes strain, which is found in rodents in Argentina and is known to spread between people. It starts with flu-like symptoms but can progress to severe breathing problems that require intensive hospital care. The mortality rate for those infected can be as high as 50%. For now, the WHO is working with officials in Argentina and on-board the ship to reconstruct the movements and contacts of infected passengers in the eight weeks before they developed symptoms, the longest known incubation period for the virus. Samples retrieved from patients are also being sequenced to help determine where the infection began and how it spread. Such insights will help authorities better understand the risk to the remaining passengers. With the exception of those who have fallen ill, who were taken off the ship at Cape Verde and airlifted to European hospitals, nobody has been allowed to disembark since May 3rd. (One infected passenger disembarked before falling ill and took himself to a hospital in Zurich.) In the meantime, the WHO has brokered a plan for those on board to leave the ship at the Canary Islands, a territory of Spain, when the ship arrives there around May 10th. The country’s health minister said that Spanish passengers would be quarantined at a military hospital, whereas other nationals will be repatriated if they are asymptomatic. The details of these plans were still being hashed out. But if they are carried out responsibly, the chances of a global health disaster are, thankfully, low. ■

Correction: An earlier version of this story incorrectly named the ship’s doctor as one of the patients. We regret the error Curious about the world? To enjoy our mind-expanding science coverage, sign up to Simply Science, our weekly subscriber-only newsletter. This article was downloaded by zlibrary from https://www.economist.com//science-and-technology/2026/05/06/how-worried-should- you-be-about-hantavirus

Science & technology | Meet the peptideins The human genome encodes for a new category of molecule They may be useful targets for future drugs May 7th 2026 In science, whether an anomaly is insignificant or the basis for a promising new field of study can boil down to the catchiness of its name. Pick the wrong one, and conferences are hard to organise and funding shrivels up. But pick the right one, and the publicity takes care of itself. In that spirit, say hello to peptideins: a newly named class of molecules found within human cells that are similar to proteins but smaller and with vaguer purposes. As the authors of the paper that named them, published in the journal Nature this week, point out, they may still be important. Sebastiaan van Heesch, a protein specialist at the Princess Máxima Centre in the Netherlands who co-led the new study, has said that peptideins might

“unlock new insights and drug targets across human biology”, potentially assisting with the development of immunotherapies and vaccines against cancer. What is certain for now is that they complicate the conventional picture of how cells work. Simply put, enzymes within a cell are thought to copy strands of DNA into molecules of RNA. These in turn are used as blueprints to make chains of amino acids known as peptides. (This second process, known as translation, is largely handled by cellular components called ribosomes.) Proteins are a loosely defined subset of peptides: those that are of a certain size; have a known function; and can be found across several species. There are about 19,500 recognised human proteins, each of which shares its name with its functional gene—the stretch of DNA responsible for making it. Many are important. The p53 protein responds to DNA damage and either pauses cell growth and division or triggers cell death as a way to suppress cancer. Insulin, a hormone, is a protein that regulates blood sugar by instructing cells to absorb glucose. For years cell biologists focused on the bits of the genome that were known to code for the proteins, with the rest dismissed as junk. But better experimental tools have revealed cracks in this simple picture and shown that DNA has valuable functions beyond protein manufacture. A technique known as ribosome profiling, for example, which can reveal the precise spot on an RNA strand where translation is under way, allowed researchers to map exactly where the cell’s protein-making machinery is active. It revealed that translation might be happening in genome regions far from known genes. At the same time, increasingly sensitive mass spectrometry experiments allowed for ever smaller molecules to be spotted in cell samples. Together, these techniques shifted the focus towards sections of DNA capable of making molecules that look like miniature proteins. The dark proteome, as this collection of “microproteins” is known, has befuddled scientists even as it has grown. Though they were once dismissed as unworthy of attention, recent studies have suggested that microproteins could be concealing drivers of disease. They could be fruitful targets for future drugs to aim at.