Anthropic, the company behind Claude AI, has identified five cases in which users employed its artificial intelligence models for research that could potentially support biological weapons development.
The company disclosed this in its latest threat intelligence report published on Thursday, saying the cases formed part of a broader range of malicious or potentially harmful activities detected and disrupted on its platform between December 2025 and August 2026.
Anthropic flagged five Claude AI use cases involving research areas including chikungunya virus, highly pathogenic avian influenza, orthopoxviruses, venom peptides and toxins.
The company said it had banned the accounts involved and incorporated findings from its investigations into its safeguards, enforcement mechanisms and threat intelligence systems.
Anthropic, however, stressed that it was not claiming that the researchers involved intended to cause harm.
“We do not assert that they intended harm, and identifying them or their labs could expose them to harm,” the company said, adding that it would not disclose the names of the institutions, countries or specific biological agents and techniques involved.
The company said the cases demonstrated the difficulty of distinguishing legitimate scientific research from activities that could potentially be used for harmful purposes.
“Biological misuse is one of the most serious risks of frontier AI models,” Anthropic said, warning that advanced AI systems could potentially be used to make existing pathogens more dangerous or facilitate the creation of new ones.
It added that the same scientific information could also be used for beneficial purposes, including developing vaccines and treatments for diseases.
In the first case, Anthropic said its biological safety classifier blocked a request to help prepare a grant application involving gain-of-function research on chikungunya virus.
The company said a subsequent investigation revealed that a reseller platform was being used to provide Claude access to researchers in regions where Anthropic does not offer its services. The platform allegedly redirected requests rejected by Claude to other AI models with less restrictive safeguards.
Anthropic said it banned the accounts involved, worked with partners to dismantle the relay networks and shared its findings with AI companies and government authorities.
In a second case, the company said a researcher outside the United States used Claude over several weeks while planning research involving highly pathogenic avian influenza and its adaptation to mammals.
According to Anthropic, the researcher used Claude for study planning, data analysis, interpretation and experiment prioritisation, as well as editorial assistance.
The company said its safeguards restricted the researcher to weaker Claude models and that the AI’s role was largely limited to clerical assistance, study ideation and research design.
Anthropic said the case nevertheless provided evidence of active research programmes involving pathogens with enhanced pandemic potential.
A third case involved a grant application for orthopoxvirus research at a state-associated infectious disease laboratory. Anthropic said a reseller relay serving more than a dozen customers was used to draft the application.
The company said the application focused on research into how orthopoxviruses interact with the immune system and was produced with Claude Opus 5 in about an hour.
The fourth and fifth cases involved research into venoms and toxins.
In one case, Anthropic said a researcher used Claude to develop an atlas of venom toxin peptides and a generative system for optimising toxin characteristics, with the stated aim of developing therapeutic molecules.
In another, a researcher used Claude in projects involving the computational redesign of toxins under a national public research programme. Anthropic said the researcher deliberately obscured the identities of some toxins and viral proteins in progress reports.
The company said both accounts were banned in May 2026 for violating its Supported Regions Policy.
Anthropic said the cases underscored the challenges involved in regulating AI-assisted biological research because many applications are inherently “dual use”, meaning they can have both beneficial and harmful applications.
The company cautioned against interpreting its findings as evidence that Claude is currently enabling imminent biological threats.
“We take these cases as evidence not of the imminence of biological threats currently uplifted by Claude, but rather as evidence that significant dual-use research efforts are associated with state actors of concern,” Anthropic said.
The report forms part of a broader disclosure by Anthropic concerning misuse of its AI systems.
The company said its investigations over the past eight months also uncovered cyber operations, surveillance, influence operations, scams and fraud, conventional weapons development and attempts to illicitly distil its models.
The report comes amid growing concern over the potential risks posed by increasingly capable AI systems. Former Anthropic researcher Jacob Coxon recently resigned from the company and publicly warned about the possibility of advanced AI systems posing catastrophic risks.
Anthropic’s latest report also disclosed a Russia-linked cyber espionage operation in which Claude was allegedly used to automate parts of attacks against organisations in Ukraine and Europe.
