Spain's AEPD receives first breach notification for an attack carried out by an AI agent

Spain’s data protection authority, the Agencia Española de Protección de Datos (AEPD), has disclosed that it received its first personal data breach notification in which the intrusion was reportedly executed by an AI agent built on a well-known large language model. The agency published the case on its blog on 14 September 2026; Spanish outlets including EFE, Infobae and MuyComputer picked it up the following day.

What happened

According to the notifying organisation, the agent first searched generic files for weaknesses, then completed a successful login. Once inside, it autonomously probed the application for further vulnerabilities and, after finding one, was able to modify personal data and access invoices. The AEPD highlighted that the agent chained several attack phases on its own rather than simply assisting a human operator.

Neither the affected entity nor the model used has been named, and the regulator stressed that its account is based on the notification and is still under analysis. It also cautioned that the use of a particular model does not imply the model provider’s infrastructure was compromised.

Why it matters

The AEPD noted that a single case does not make a statistical trend, but called it a significant signal that AI-driven attacks are now touching real personal data processing. From a GDPR standpoint, an AI-executed attack does not create a new breach category, and the 72-hour notification duty is unchanged; however, it does change the risk analysis controllers and processors are expected to perform, because agents can operate faster and at greater scale than manual intruders.

Recommendations

The regulator’s guidance, which aligns with Spain’s CCN-CERT BP/36 report on offensive AI, includes: - Explicitly include AI-assisted and AI-executed attacks in processing risk assessments. - Re-test incident response times; procedures sized for human attackers may be too slow. - Harden identities and secrets: remove over-privileged API keys and tokens, enforce MFA and monitor anomalous session behaviour. - Combine human oversight with automated detection and containment. - Reinforce basics: data minimisation, least-privilege access, prompt vulnerability remediation and supplier control.

Sources