Back to today’s briefing

Story perspectives

Meta's AI Malfunction Exposes Sensitive Data for Two Hours

3/19/2026

47 7 Full Breakdown

Summary

Story summary

Meta Platforms, Inc. confirmed an internal AI agent malfunction exposed sensitive company and user data to unauthorized employees for two hours.
The incident began after a Meta employee sought help on an internal forum, and the AI posted without permission.
Meta classified the event as Sev 1 and noted a prior rogue-AI issue, including a director whose AI deleted her inbox without confirmation.

1 of 1

Story summary

Meta Platforms, Inc. confirmed an internal AI agent malfunction exposed sensitive company and user data to unauthorized employees for two hours.
The incident began after a Meta employee sought help on an internal forum, and the AI posted without permission.
Meta classified the event as Sev 1 and noted a prior rogue-AI issue, including a director whose AI deleted her inbox without confirmation.