AI Safety Alert: Anthropic Blocks Attempts to Use Claude for Biological Weapons Research

Claims that AI threatens to wipe out the human race are increasing.

Anthropic has successfully disrupted multiple attempts to exploit its artificial intelligence models for research that could potentially aid in the development of biological weapons, the United States technology company announced.

The firm detailed these incidents in a report published on Thursday, highlighting instances of misuse involving its models—including the chatbot Claude—over the preceding eight months, amid growing anxieties surrounding the risks of rapid artificial intelligence advancement.

Anthropic explained that it was often difficult to distinguish between legitimate and malicious inquiries, noting that scientific exploration aimed at developing vaccines can sometimes overlap with research required to engineer hazardous pathogens. In ambiguous situations, the company chose to err on the side of caution.

Highlighting a specific incident from May, Anthropic noted that a scientist requested assistance with a grant application to engineer mutations to chikungunya—a mosquito-borne virus comparable to malaria—to increase its harmful effects in live animals. Company officials found this case particularly alarming because the requested work appeared intended for a military research institute, prompting Anthropic to ban the associated accounts.

“What we don’t know is if the research was meant to be weaponised,” Jacob Klein, Anthropic’s head of threat intelligence, stated to The New York Times.

Andrew Weber of the Council on Strategic Risks, who reviewed the report prior to its publication, told the newspaper that access to such advanced models should be restricted to trusted researchers.

According to Anthropic, evaluations conducted on its older models last year indicated they were incapable of providing meaningful assistance with dangerous biological research, whereas newer iterations are capable of executing complex scientific tasks.

The same report also outlined attempts to utilize Claude for missile projects and espionage.

Just two days prior to the report's release, a researcher resigned from Anthropic, warning that the competitive rush to develop artificial intelligence could ultimately destroy humanity.

Meanwhile, President Donald Trump dismissed mounting concerns regarding the technology late on Thursday. Addressing reporters, President Donald Trump emphasized his fear that the United States would find itself in a “very bad position” if it failed to secure leadership in the global AI race. When asked if he harbored fears that artificial intelligence could trigger human extinction, he responded, “No, I don’t have any.”