Every AI agent followed the rules, and the data still leaked

IBM Consulting highlights how AI agents, even when following all access controls, can cause data leaks due to gaps in identity propagation and integration across systems, risking GDPR fines and costly data breaches. The 2026 IBM Cost of a Data Breach Report shows AI-related breaches are rising, with identity failures as the primary root cause, underscoring the urgent need for enterprises to implement comprehensive identity and access controls across legacy environments.

This Wire brief sits within Fusion42's coverage of AI Agents, Identity & Access and Cybersecurity.