{"slug": "someone-used-claude-to-build-a-potential-bioweapon-the-real-threat-is-much", "title": "Someone used Claude to build a potential bioweapon. The real threat is much deeper", "summary": "Anthropic reported in its September 2026 threat intelligence report that anonymous scientists attempted to use its flagship Claude model for potentially dangerous research, including a grant application for gain-of-function work on the mosquito-borne Chikungunya virus and the development of novel venoms and toxins; the company said it blocked the requests and found no evidence the scientists intended harm. Anthropic warned that because chikungunya circulates naturally, a deliberate release as a bioweapon would be difficult to distinguish from a natural outbreak, and that dual-use toxin research poses a grave challenge for frontier model developers. The incidents highlight the growing risk that increasingly capable LLMs could assist in creating lethal pathogens or toxins.", "body_md": "Today’s frontier [AI](https://www.fastcompany.com/section/artificial-intelligence) models know everything–how to safely thaw a frozen chicken breast, re-shingle your roof, and treat your dog’s ragweed allergies, if my recent chat history is any indication.\n\nApparently, they also know how to create fiendishly [deadly bioweapons](https://www.fastcompany.com/90968614/rand-report-ai-dangers-bioweapons).\n\nThat’s according to a [recent announcement by Anthropic](https://www.anthropic.com/threat-intelligence-report-september-2026). According to the company, anonymous scientists attempted to use [Anthropic’s flagship Claude](https://www.fastcompany.com/91609568/claude-anthropics-ai-model-helping-develop-next-version-itself) model to conduct research that could have turned deadly.\n\nAnthropic blocked their efforts this time, and there’s no evidence that the scientists were actually trying to cause harm.\n\nBut as [frontier models get more powerful](https://www.fastcompany.com/91605458/ai-doomsday-dictionary) and better at science, the threat of an LLM imagining a truly lethal new virus or bacteria will only increase.\n\nAccording to Anthropic’s report, the potentially dangerous research that the company blocked varied tremendously.\n\nIn one instance, scientists reportedly asked Claude to assist with a grant application for so-called “gain of function” research on the mosquito-borne Chikungunya virus.\n\nGain of function research involves [deliberately making a pathogen more deadly or easier to spread](https://www.ncbi.nlm.nih.gov/books/NBK285579/), in order to learn about how natural versions of the pathogen infect their hosts.\n\nSuch research can have legitimate scientific purposes, and it’s possible that the scientists (who Anthropic did not identify) were indeed trying to write an application to fund their legitimate work.\n\nBut asking for help with something nefarious in the guise of a request for some other form of assistance (like a grant application) is a classic example of “jailbreaking”—methods that can trick an LLM into providing information or guidance that its guardrails should prevent.\n\nIn this case, Anthropic pointed out, a deadlier virus could have incredibly lethal consequences.\n\n“Because chikungunya circulates naturally, a deliberate release (as part of a bioweapon) would be difficult to distinguish from a natural outbreak,” Anthropic says in its report.\n\nIn other words, a natural virus that already circulates in mosquitos could be made far deadlier overnight, and scientists would have no idea the “gain of function” was caused by AI’s meddling.\n\nIn another incident, Anthropic says that scientists tried to use Claude to develop “novel venoms and toxins.”\n\nThis vein of research is especially alarming because it involves a “dual use” technology.\n\nAs Anthropic points out, there are plenty of legitimate reasons to develop new toxins–Botox, for example, comes from a deadly bacteria toxin, but is also used ”to treat migraines, spasticity, and wrinkles.”\n\nBut the same research that creates a more effective way to inject away your crow’s feet could also be used to create a more powerful tool for covert assassinations. Indeed, ricin (another natural toxin) was famously embedded in a weapon [disguised as an umbrella and used to assassinate a journalist during the Cold War.](https://www.bbc.com/culture/article/20260904-the-poison-tipped-umbrella-that-pointed-to-a-kgb-murder)\n\nDual use technology presents an especially grave challenge for firms like Anthropic. The company doesn’t want to squash potentially promising scientific research (or drive big customers to a competitor like OpenAI) by shutting legitimate inquiries down.\n\nBut if Anthropic misses a nefarious use of tech which could also have positive uses, the results could be dire–and deadly.\n\nIn these instances, Anthropic says that it successfully blocked potentially harmful requests. But there’s no way to know how many similar requests made it past the company’s filters, or its technology’s guardrails.\n\nThe specific instances shared in Anthropic’s report also highlight the challenges of policing LLMs. It’s easy enough to block a request like “make me a stronger version of COVID.”\n\nIt’s far harder to know when potentially valuable scientific inquiries are actually masking an attempt to design a deadly toxin or develop a highly-contagious new pathogen.\n\nBefore–as Anthropic itself admits–the company could hide behind the fact that its models simply weren’t that good at science. The design for a deadly new pathogen won’t do anyone any harm if the biology doesn’t work.\n\nAs frontier models get better at scientific reasoning, though, the threat will only increase. And while highly-controlled models like Claude are easier to lock down, open-weight LLMs are only a few months behind today’s top frontier models, and often include almost no guardrails.\n\nAs LLMs advance, we may reach a point where the only thing preventing bad actors from developing powerful new pathogens is their desire not to poison the world.\n\nFor anyone with a dim view of human nature and a desire to keep the species humming, that will be little comfort.", "url": "https://wpnews.pro/news/someone-used-claude-to-build-a-potential-bioweapon-the-real-threat-is-much", "canonical_source": "https://www.fastcompany.com/91608269/someone-used-claude-build-potential-bioweapon-real-threat-much-deeper", "published_at": "2026-09-19 11:00:00+00:00", "updated_at": "2026-09-19 11:23:05.614235+00:00", "lang": "en", "topics": ["ai-safety", "ai-policy", "large-language-models", "artificial-intelligence", "ai-ethics"], "entities": ["Anthropic", "Claude", "Chikungunya virus", "OpenAI", "Botox", "ricin"], "alternates": {"html": "https://wpnews.pro/news/someone-used-claude-to-build-a-potential-bioweapon-the-real-threat-is-much", "markdown": "https://wpnews.pro/news/someone-used-claude-to-build-a-potential-bioweapon-the-real-threat-is-much.md", "text": "https://wpnews.pro/news/someone-used-claude-to-build-a-potential-bioweapon-the-real-threat-is-much.txt", "jsonld": "https://wpnews.pro/news/someone-used-claude-to-build-a-potential-bioweapon-the-real-threat-is-much.jsonld"}}