Anthropic reveals hacker used Claude to target French far

A lone hacker attempted to breach around 40 organizations linked to the French far right, stealing significant amounts of personal data in several cases, the artificial intelligence company Anthropic revealed in its latest security report published Thursday, September 10.

In the spring, a French-speaking hacker used Anthropic's AI service, Claude, to refine a targeting and hacking infrastructure that was then deployed against "European political parties, media, think-tanks and the SaaS [technical service] providers used by these organizations," according to the company. The report did not specify the names of the targets, but technical data suggested the vast majority were French and associated with the far right. The victims included at least one French political party, a political training institute, several news websites including the far-right magazine Frontières, and a discussion forum linked to a podcast.

Of "42 tracked target entities," the hacker "gained internal access to at least 14," wrote Anthropic. Technically skilled, the hacker "exfiltrated an estimated 12 to 26 GB of database dumps, including information on political party donors and member records." In the case of Frontières, Anthropic's information appears to match a cyberattack that targeted the magazine's website in the spring, using spyware injected into its comments section. Since then, the site has disabled its comments feature.

You have 56.59% of this article left to read. The rest is for subscribers only.