Anthropic discloses 5 blocked bioweapon research cases involving researchers from Russia, China, Iran
Anthropic disclosed on September 11 five bioweapon research cases it blocked this year, including researchers from Russia, China, and Iran, with one case involving avian influenza experiment planning.
Anthropic disclosed in a report on September 11 five cases in which the company blocked bioweapon-related research this year. The affected users were all from regions where Anthropic restricts model access, including Russia, China, and Iran.
One case involved a researcher from a restricted region who spent weeks planning avian influenza experiments using Claude, working around Anthropic's regional access rules until safety filters restricted the work to the company's weakest models. Anthropic said it could not confirm whether the researcher intended harm and noted that the information required to design a bioweapon can also be used to design a vaccine.
The report also expanded Anthropic's earlier claims that seven China-based labs were distilling Claude, explicitly naming Moonshot AI and DeepSeek as using increasingly sophisticated methods to circumvent its defenses and harvest the capabilities of U.S. frontier models.
The disclosure comes days after former Anthropic researcher Jacob Coxon publicly resigned. Coxon posted on X that he earnestly believes AI could kill us all by the end of the decade. Anthropic scientist Evan Hubinger responded that Coxon's warning was not overstated.