AEPD discloses first breach notification tied to an autonomous AI agent

Spain's data protection authority, the AEPD, disclosed on September 14, 2026 that it had received its first reported notification of a personal data breach allegedly carried out by an AI agent. The regulator said the agent, built on a widely known large language model, located a weakness, logged into a system and then altered records and reviewed billing documents.

What happened

The AEPD described a breach report submitted by an unnamed organization, in which an unidentified hacker used a large language model against corporate systems. According to Dark Reading's account, the AI appears to have discovered and exploited a set of loose credentials, plus an enterprise application vulnerability, allowing its proprietor to modify personal data records and access corporate invoices. AEPD did not immediately respond to a Reuters request for comment, and the regulator has not named the organization or the model involved.

Impact

Personal data was modified and invoices were accessed, according to the organization's own notification, and the Spanish watchdog did not say when it would finish reviewing the reported breach. No figures on the number of records affected have been disclosed, and both the victim organization and the AI model remain unidentified.

Response

The agency said that the alleged breach was reported to it by the affected organization, and it stressed the case still needs analysis. AEPD also clarified that using a particular model did not mean either the model itself or its provider's infrastructure was compromised, nor that the tool was built for abuse. In its security alert, AEPD highlighted the importance of securing digital identities and credentials against offensive AI. It also recommended that organizations review and potentially speed up their incident response times, and support manual intervention into data breaches with.

Lessons for security leaders

The agency said the case was relevant because a third party allegedly used an AI agent to carry out multiple stages of an attack with limited human intervention. The agency said AI does not create new threats. Controllers, processors and data protection officers must prepare for a scenario in which the. Security leaders should treat the case as a prompt to tighten credential hygiene, test detection and response at machine speed, and ensure AI-driven attack paths are included in existing risk assessments.