Meta’s sensitive information leaked to employees after engineer seeks AI agent who goes rogue

Facebook-parent Meta is reportedly having trouble with rogue AI agents. According to a report by The Information, a Meta AI agent malfunctioned, exposing sensitive company and user data to employees who did not have permission to access it. Per the report, the employee posted on the company’s internal forum asking for help with a technical question. Another engineer asked the AI agent to help analyze the question. However, the agent posted a response without asking the engineer for permission to share it.

What happened next was concerning. The engineer who asked the question ended up taking actions based on the AI agent’s guidance, which as per the report did not give good advice.