Meta is having trouble with rogue AI agents
By Amanda Silberling
Published on March 18, 2026.
An AI agent went rogue at Meta, inadvertently exposing sensitive company and user data to employees who did not have permission to access. The incident was deemed a "Sev.
" by the company, which also noted the potential for agentic AI. The company recently purchased Moltbook, a social media site for interacting with agents.
Read Original Article