{"slug": "anthropic-disrupts-attempts-to-use-claude-ai-for-biological-weapons-research", "title": "Anthropic disrupts attempts to use Claude AI for biological weapons research", "summary": "Anthropic disrupted multiple attempts to use its Claude models for high-risk biological research between late 2025 and mid-2026, according to a threat intelligence report the AI safety lab published in September 2026. The report documents five case studies, including a scientist linked to a military research facility who sought Claude's help drafting a gain-of-function grant proposal for the mosquito-borne chikungunya virus, a user who drafted an orthopoxvirus immune-evasion proposal on Claude Opus in roughly one hour via third-party API reseller relays, and state-linked attempts to optimize biological venoms and toxins. Anthropic said it banned the offending accounts and deployed layered mitigations, including hard refusals, while noting that definitive malicious intent is hard to establish because biochemical queries used to model vaccine targets mirror those used to engineer dangerous pathogens.", "body_md": "As frontier artificial intelligence models grow increasingly capable of assisting complex scientific workflows, the boundary between legitimate healthcare research and dangerous biological synthesis has become an urgent [national security](/tech/72706/doj-urges-court-to-toss-data-center-pollution-case-claiming-national-security-threat) frontier. In a comprehensive threat intelligence briefing published by [Anthropic](/tech/72395/anthropic-ceo-says-ai-growth-is-exponential-anthropic-research-says-otherwise), the AI safety lab revealed it disrupted multiple attempts to harness its [Claude models](/tech/75315/claude-ai-content-watermarking-explained-what-it-means-for-text-code-and-image-files) for high-risk biological research between late 2025 and mid-2026. The documented incidents, which involved working scientists and state-backed entities attempting to bypass model safeguards, underscore the emerging dual-use challenge facing frontier [AI developers](/tech/67108/spotify-says-its-top-developers-havent-written-a-line-of-code-since-december-new-internal-ai-has-mad) as foundational models acquire the capacity to accelerate pathogen engineering and conventional arms software.\n\nWe're publishing our most detailed threat intelligence report to date.\n\nIt covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them.\n\nWe disrupted every operation in the report,…\n\n[September 10, 2026](https://x.com/AnthropicAI/status/2098097512544444447?ref_src=twsrc%5Etfw)\n\n### **Dual-Use ambiguity and pathogen studies**\n\nAccording to Anthropic's report, biological misuse represents one of the most severe threat categories confronting [frontier model](/tech/75887/jacob-coxon-resigns-anthropic-citing-ai-dangers-decoding-the-viral-x-thread) developers. The lab documented five specific case studies where users sought assistance that could support biological weapons pipelines:\n\n**Gain-of-Function Grant Drafting:** In one case, a scientist linked to a military research facility sought Claude's help drafting a grant proposal for gain-of-function research on the mosquito-borne chikungunya virus, aiming to engineer mutations that increased transmission in live animal hosts.\n\n**Orthopoxvirus Immune Evasion:** Another user utilized third-party API reseller relays to draft an orthopoxvirus immune-evasion proposal on [Claude Opus](/tech/74218/anthropic-release-claude-opus-5-its-safest-model-yet) in approximately one hour.\n\n**Avian Influenza and Toxin Redesign:** Additional cases involved mammalian-adaptation experiments on avian influenza, as well as state-linked attempts to optimize biological venoms and toxins.\n\n[Anthropic](/tech/75879/anthropic-researcher-quits-says-ai-could-kill-us-all-by-the-end-of-the-decade) emphasized that while the activities posed severe safety risks, establishing definitive malicious intent is difficult because the biochemical queries used to model vaccine targets directly mirror those used to engineer dangerous pathogens.\n\n🚨 THE AI THREAT EVERYONE FEARED IS ALREADY HAPPENING\n\nAnthropic just released a major threat intelligence report revealing how state-linked groups, hackers and criminals are weaponizing AI.\n\nThe report covers malicious activity disrupted between December 2025 and August 2026.… [pic.twitter.com/TApfOpLIwA](https://t.co/TApfOpLIwA)\n\n[September 11, 2026](https://x.com/BullTheoryio/status/2098375881576403154?ref_src=twsrc%5Etfw)\n\n### **Evasion tactics and safety interventions**\n\nThreat actors utilized multiple circumvention techniques, including routing requests through third-party [API](/tech/56709/the-x-api-costs-developers-42k-per-month-now-x-wants-a-cut-of-their-revenue-instead) relays, operating from embargoed jurisdictions, and chaining automated fallbacks to competitor models when Claude refused prompts.\n\nIn response, Anthropic banned the offending accounts and deployed layered mitigations, including hard refusals via specialized biological safety classifiers, [automated model](/tech/75832/microsoft-engineer-declares-typing-code-is-over-how-ai-is-transforming-software-development) downgrades to less capable tiers, and proactive institutional sweeps.\n\nAnthropic’s public disclosure signals that passive keyword filtering is no longer enough to manage frontier model risks. As autonomous AI agents lower the technical barrier for biochemical experimentation, AI developers must transition toward verified access credentials and continuous behavioral telemetry to prevent foundational models from accelerating biological proliferation.\n\n##### *(Feature image credits to Andrey Rudakov/Bloomberg via Getty Images)*\n\n**Read More:** [iPhone 18 Pro Max vs. iPhone 17 Pro Max: Everything Apple changed](/tech/75901/iphone-18-pro-max-vs-iphone-17-pro-max-everything-apple-changed)", "url": "https://wpnews.pro/news/anthropic-disrupts-attempts-to-use-claude-ai-for-biological-weapons-research", "canonical_source": "https://me.mashable.com/tech/75960/anthropic-disrupts-attempts-to-use-claude-ai-for-biological-weapons-research", "published_at": "2026-09-11 12:21:42+00:00", "updated_at": "2026-09-11 12:38:27.474558+00:00", "lang": "en", "topics": ["ai-safety", "ai-policy", "artificial-intelligence", "large-language-models", "ai-ethics"], "entities": ["Anthropic", "Claude", "Claude Opus", "chikungunya virus", "orthopoxvirus", "avian influenza"], "alternates": {"html": "https://wpnews.pro/news/anthropic-disrupts-attempts-to-use-claude-ai-for-biological-weapons-research", "markdown": "https://wpnews.pro/news/anthropic-disrupts-attempts-to-use-claude-ai-for-biological-weapons-research.md", "text": "https://wpnews.pro/news/anthropic-disrupts-attempts-to-use-claude-ai-for-biological-weapons-research.txt", "jsonld": "https://wpnews.pro/news/anthropic-disrupts-attempts-to-use-claude-ai-for-biological-weapons-research.jsonld"}}