Anthropic said on Thursday that it had identified and disrupted attempts to use its AI model for "malicious activity" that could support the development of biological weapons. The company’s latest threat intelligence report lands in the middle of a familiar corporate ritual: first the warning, then the product pitch, then the promise that the same firm that built the machine can keep the machine from biting back.
Corporate Guardians, State Clients
Anthropic said the report is the latest in a growing series of warnings from AI researchers and industry insiders about the technology's potential risks to humanity. Those warnings have already spilled into the political arena. US Senator Bernie Sanders demanded a pause on advanced AI development and a ban on artificial superintelligence, while President Donald Trump rejected that line of alarm on Thursday, saying he was concerned "if we don't win AI, we're going to be put in a very bad position". The language is blunt. The logic is the same old race: build faster, regulate later, and call it strategy.
Anthropic said such reports have become a regular feature across the AI industry as firms seek to demonstrate how they identify and disrupt attempts to misuse their models. Google wrote on Tuesday that a person had attempted to use its AI tool, Gemini, to obtain a "complete, step-by-step technical guide for synthesizing weaponised biological agents". The companies are now competing to prove they can spot the danger they helped put into circulation. That’s the market speaking in the language of public safety.
Anthropic said Claude, its AI model, was also used by actors linked to a Russia-based cyber espionage campaign and by an Iranian propaganda institution. The cases of concern detected over the past eight months ranged from fake dating apps and hotel Wifi scams to surveillance built to identify dissidents. Anthropic has also accused Chinese AI firms of trying to replicate Claude's capabilities. The report listed cases in which suspected state-sponsored groups, criminals, spyware vendors, state propaganda institutions and politically motivated individuals misused its technology. The list reads like a catalogue of the modern security state, except the tool at the centre of it is sold as innovation.
Weapons, Surveillance, and the Usual Clients
Anthropic said "malicious use" of its Claude Haiku, Sonnet and Opus models was disrupted between December 2025 and August 2026. It said none of the misuse cases involved Claude Fable or the powerful Mythos-class models, except for one instance of distillation, the process for training smaller AI models using larger, more expensive models. Anthropic said it had blocked scientists who used its AI in ways that could support biological weapons development. The report highlighted "five case studies of actors using our models in ways that could support biological weapons development".
The company said biological misuse is "one of the most serious risks of frontier AI model" and that without the correct safeguards such capabilities "could have catastrophic consequences". Anthropic said: "The same information that can be used to develop a biological weapon could also be used to develop, for example, a vaccine or a cure for a disease." Jacob Klein, the head of threat intelligence at Anthropic, told the New York Times it was "an incredibly nuanced situation". He said: "You are not seeing someone in a comic book kind of way say, 'Hey, I want to build a biological weapon to kill everybody,'". The quote does a lot of work. It softens the horror into a management problem, as if catastrophe only counts when it arrives wearing a cartoon villain mask.
The report also noted six cases where Claude was used "to develop software for conventional weapons, including firearms, missiles, armed drones, bombs, and other munitions, as well as the targeting and control systems that operate them". Anthropic said cybercriminals and state-backed hackers have increasingly used its technology to assist their operations. Hacking group ShinyHunters, as well as China-based labs, were among those named in the report. The report said a hacking group whose work is consistent with the Russia-based Midnight Blizzard allegedly used AI to build a system that automatically detected when its malware was flagged by security defences and rewrote code until it evaded detection.
The Safety Industry Protects Itself
The California-based company said it had incorporated its findings into its processes "to better prevent, detect, and disrupt these activities in the future". Anthropic said it had shared intelligence with authorities and industry partners where appropriate. That’s the loop: the company detects the misuse, reports the misuse, shares the misuse with authorities, and then asks for trust because it has now learned something useful from the misuse.
The report, the company's first this year, comes after a top safety researcher at Anthropic warned AI is advancing so quickly he believes there is a greater than 10% chance it "could kill all humans" within the next decade. In an article published on 6 September, OpenAI chief scientist Jakub Pachocki called for the industry to implement "voluntary slowdowns" until safeguards are set. Pachocki said: "I am concerned no one is prepared for the consequences of a continued rapid rise in machine intelligence," adding that OpenAI, which makes ChatGPT, will continue its work on building safeguards, though broader interventions are required.
The warning prompted an open letter to UK Prime Minister Andy Burnham calling for a new multinational treaty for the safe development of AI and a call for governments to collaborate about what a treaty based on the development of superintelligence should look like. In the US, Democratic lawmaker Bernie Sanders has introduced legislation to ban AI superintelligence and temporarily pause advanced AI development. Sanders said on Thursday on BBC's Newsnight: "When scientists tell you there is a chance, a chance that it could have a cataclysmic impact on humanity, you've got be a moron not to say, slow it down." The political class can hear the alarm. The machine keeps running anyway.