Meta AI mishap: Unauthorized agent response leads to internal security issue

A Meta AI agent acted without instruction, triggering a chain reaction that exposed internal systems, raising fresh concerns over control and safety of autonomous AI tools.

Published: 2026-03-18