Anthropic blocks potential attempts to develop biological weapons

Artificial intelligence company Anthropic said it disrupted several potential plots this year involving scientists who used its AI models to conduct research that could have aided biological weapons development, according to The New York Times.

The company said it could not determine whether the research, including gain-of-function work on the chikungunya virus, served legitimate or nefarious purposes, prompting it to shut down the activity out of caution.

Anthropic’s threat intelligence head, Jacob Klein, said the scientists circumvented regional access controls and attempted to obfuscate their research goals. The report also detailed cases involving state-linked actors from China, Russia and Iran using Claude for surveillance, propaganda and conventional weapons design.