Cyber Awareness

Anthropic Says It Stopped Attempts to Misuse AI for Bioweapons Research

By Today News Hours Editorial Desk

Anthropic has revealed it identified and shut down attempts to use its artificial intelligence models for malicious purposes, including activity that could have supported the development of biological weapons. The disclosure, made in a threat intelligence report published Thursday, adds to a growing list of warnings from AI researchers and industry figures about the technology's potential dangers.

These warnings have fueled calls for action. US Senator Bernie Sanders has demanded a halt to advanced AI development and a ban on artificial superintelligence. President Donald Trump, however, has dismissed such concerns, stating Thursday that he worries "if we don't win AI, we're going to be put in a very bad position."

Anthropic's report is part of a broader trend of AI companies releasing similar findings. Google reported Tuesday that someone had attempted to use its Gemini AI tool to obtain a "complete, step-by-step technical guide for synthesizing weaponised biological agents."

According to Anthropic, its Claude AI model was exploited by actors linked to a Russia-based cyber espionage campaign and by an Iranian propaganda organization. Over the past eight months, the company detected a range of concerning activities, including fake dating apps, hotel WiFi scams, and surveillance tools designed to track dissidents. Anthropic also accused Chinese AI firms of attempting to copy Claude's capabilities.

The report details cases where suspected state-sponsored groups, criminal organizations, spyware vendors, propaganda institutions, and politically motivated individuals misused its technology. Anthropic said it disrupted malicious use of its Claude Haiku, Sonnet, and Opus models between December 2025 and August 2026. Its more advanced Claude Fable and Mythos-class models were not implicated, aside from one case involving distillation—a process used to train smaller AI systems using larger, costlier ones.

Among the findings, Anthropic identified five case studies of actors using its models in ways that could support biological weapons development, calling this one of the most serious risks posed by advanced AI. The company noted that without proper safeguards, such capabilities could lead to catastrophic outcomes, though it acknowledged the dual-use nature of the information involved. "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," the company stated.

Jacob Klein, Anthropic's head of threat intelligence, told the New York Times the situation is highly nuanced, noting that misuse doesn't typically involve someone explicitly stating an intent to cause mass harm.

The report also documented six instances where Claude was used to develop software for conventional weapons, including firearms, missiles, armed drones, bombs, and related targeting and control systems.

Additionally, the report highlighted growing use of AI by cybercriminals and state-backed hackers. The hacking group ShinyHunters and China-based labs were named, along with a group whose activity aligns with the Russia-linked Midnight Blizzard operation. That group allegedly used AI to build a system capable of detecting when its malware was flagged by security software and automatically rewriting the code to evade detection.

Anthropic, based in California, said it has incorporated these findings into its internal processes to better prevent, detect, and disrupt future misuse, and has shared relevant intelligence with authorities and industry partners.

This marks the company's first such report this year. It follows a warning from a senior Anthropic safety researcher who said AI is progressing so rapidly that there is more than a 10% chance it could kill all of humanity within the next decade.

Separately, OpenAI's chief scientist Jakub Pachocki wrote on September 6 calling for the industry to adopt voluntary slowdowns until adequate safeguards are established, expressing concern that no one is prepared for the consequences of rapidly advancing machine intelligence. He said OpenAI, the maker of ChatGPT, would continue developing safeguards but that broader industry-wide action is necessary.

These warnings prompted an open letter to UK Prime Minister Andy Burnham calling for a new multinational treaty to ensure the safe development of AI, along with calls for global cooperation on regulating superintelligence.

In the US, Senator Sanders has introduced legislation seeking to ban AI superintelligence and impose a temporary pause on advanced AI development. Speaking on BBC's Newsnight Thursday, he said, "When scientists tell you there is a chance, a chance that it could have a cataclysmic impact on humanity, you've got to be a moron not to say, slow it down."