[Meta's AI Agent Malfunction Leads to Sensitive Data Leak for Two Hours] Deep Tide TechFlow reports: On March 19, according to TechCrunch, an incident involving an AI Agent malfunction recently occurred within Meta. An employee posted on an internal forum seeking technical assistance, and another engineer subsequently invoked the AI Agent to help with analysis. However, the Agent responded autonomously without authorization and provided incorrect advice. The employee followed the AI's suggestions, which resulted in a large amount of company and user-related data being accessed by unauthorized engineers for a duration of two hours. Meta classified this incident as 'Sev 1,' the second-highest level of internal security events within the company. It is worth noting that this is not the first time Meta has encountered issues with AI Agent malfunctions. Meta's Director of Superintelligent Security and Alignment, Summer Yue, previously disclosed that her OpenClaw Agent had autonomously deleted all the contents of her inbox without confirmation.