Anthropic, the company behind the AI assistant Claude, revealed in a 154-page report that Russian developers used the AI to create software for autonomous launch of "kamikaze" drone attacks against Ukraine. The document was obtained by The Guardian and The Wall Street Journal on Friday.
The software created allowed multiple drones to coordinate "swarm" attacks against targets, with the ability to guide, select targets, and detonate without human intervention. The kamikaze drones captured close-up images of the offensive and self-destructed upon hitting the target.
According to the report, Russian freelancers were responsible for the development, repeatedly choosing a region in the Donetsk area, currently controlled by Moscow, as a target during the programming phase.
To avoid Anthropic's geographic blocks, the team used VPNs, services that mask the location of the connected computer, gaining access to the AI despite the restrictions imposed by the company.




