Editorial

Who Authorized the Robots?

Olivia Sharp 7 min read 381 views
Free
Firms deployed AI agents into nuclear plants and financial systems this week while researchers disclosed vulnerabilities in the same technology.

Someone put an AI-powered humanoid into a French nuclear facility this week. The robot, named Hoxo, will "replicate human movements" in what Orano describes as "challenging intervention environments." The company announced a four-month testing phase to validate its applications.

This happened three days after a security researcher showed he could trick Anthropic's Claude into stealing user data through a document. The attack works by hiding commands in a file. Claude reads the file, follows the instructions, packages up the user's conversation history, and ships it to an attacker's server. Up to 30 megabytes at once. Anthropic was notified. No fix …

Archive Access

This article is older than 24 hours. Create a free account to access our 7-day archive.

Share this article

Related Articles