US artificial intelligence company Anthropic said Thursday it had disrupted multiple attempts to use its models for research that could assist in the development of biological weapons, according to a company report covering the past eight months.
The firm said it could not always determine whether the inquiries were for legitimate scientific purposes, such as vaccine work, or for malicious ends, and therefore chose to act cautiously.
In one case in May, a scientist requested assistance with a grant application involving research on a mosquito-borne virus at what Anthropic identified as a military research institute. The company banned the accounts involved.
“What we don’t know is if the research was meant to be weaponised,” Jacob Klein, Anthropic’s head of threat intelligence, told The New York Times.
Andrew Weber of the Council on Strategic Risks, who reviewed the report, said access to advanced AI models should be restricted to trusted researchers.
Anthropic noted that evaluations of its older models found they could not provide meaningful help with dangerous biological work, but newer versions are capable of more complex scientific tasks.
The same report also detailed other misuse attempts involving missile-related projects and espionage.
The disclosure comes amid growing public debate over AI safety. A researcher recently left Anthropic, warning that the rapid development of the technology could pose existential risks.
US President Donald Trump, speaking late Thursday, dismissed concerns that artificial intelligence could lead to human extinction.
“No, I don’t have any,” Trump said when asked about the possibility. He stressed that the United States must lead in the AI race or risk falling into a “very bad position.”
