‘Vibe-hacking’ is now a top AI threat

"Agentic AI systems are being weaponized."

That's 1 of the archetypal lines of Anthropic's caller Threat Intelligence report, retired today, which details the wide scope of cases successful which Claude - and apt galore different starring AI agents and chatbots - are being abused.

First up: "Vibe-hacking." One blase cybercrime ringing that Anthropic says it precocious disrupted utilized Claude Code, Anthropic's AI coding agent, to extort information from astatine slightest 17 antithetic organizations astir the satellite wrong 1 month. The hacked parties included healthcare organizations, exigency services, spiritual institutions, and adjacent authorities entities.

"If you're a truthful …

Read the afloat communicative astatine The Verge.