Anthropic Says It Blocked Attempts to Use AI for Bioweapon Research

Anthropic says it blocked multiple accounts after identifying attempts to use its artificial intelligence models in ways that could potentially support the development of biological weapons.

In a report published Thursday, the company said biological misuse is among the most serious risks associated with increasingly capable AI systems.

Anthropic reviewed 30 days of activity and identified about 35 distinct research efforts involving potentially concerning activity. The company said it could not determine whether the people involved intended to cause harm or were conducting legitimate scientific research.

According to Anthropic, the cases involved working scientists who attempted to circumvent safeguards or obscure the purpose of their research. The company did not identify the researchers, institutions or countries involved.

One case involved a request for help writing a grant application for gain-of-function research examining the transmissibility and immune evasion of the chikungunya virus.

Another involved a researcher outside the United States using Claude for work involving bird flu and the adaptation of viruses to mammals.

The report also described research involving orthopoxviruses, a family that includes smallpox and mpox, as well as research into venom toxins.

Anthropic said the same capabilities that could help scientists develop medicines and treatments could potentially be misused for harmful biological research.

“As our models become increasingly capable, approaching or exceeding expert performance at challenging scientific tasks, we expect their impact will only increase, both in beneficial and potentially harmful contexts,” the company said in its report.

Anthropic said newer models have stronger safeguards designed to restrict access to a wide range of dual-use biological research requests. The company said the safeguards have shown evidence of working, but acknowledged that more robust safety systems will be needed as AI capabilities and potential threats become more sophisticated.

The disclosure comes amid broader debate over AI safety and the pace of development. AI companies have faced increasing calls from researchers and former employees for stronger safeguards and government oversight.

Anthropic said it believes AI companies should establish a lawful and verifiable way to work together on how powerful AI models are released.

Explore: NBCPalmSprings.com, where we are connecting the Valley.