Spanish DPA receives first data breach notification caused by AI agent attack
On 14 September 2026, the Spanish Data Protection Agency (AEPD) received the first notification of a personal data breach where an artificial intelligence (AI) agent executed the incident. The attacking agent initiated a search for vulnerabilities in generic files, performed a correct login, then autonomously searched for application vulnerabilities, modified personal data, and accessed invoices.
This incident, while not yet a statistical trend, signals that AI supported attacks are materialising beyond theoretical risks, affecting real personal data processing. The AEPD notes that while malicious use of artificial intelligence is not new, the emergence of AI agents introduces a qualitative change, as an agent can receive an objective, plan intermediate tasks, use tools, execute code, consult sources, interpret results, and autonomously modify its actions based on findings. Artificial intelligence does not create new threats but increases the speed, scale, and adaptability of known malicious techniques, reducing detection and containment time, a conclusion also found in the CCN CERT BP/36: Good practices against offensive AI model guide.
This notification confirms the need to explicitly incorporate AI assisted or executed attacks into data processing risk analyses, as generic references to malware or unauthorised access are insufficient given the potential for substantial changes in incident probability, speed, and scope. It also necessitates a review of response times, as procedures for manual attacks may be inadequate when an agent rapidly analyses multiple assets and adapts its behaviour. Furthermore, the incident underscores the increased importance of digital identities and credentials, as an agent obtaining an account, an API key, or a token with excessive permissions can operate at machine speed across services before anomalous behaviour is detected. The AEPD concludes that security cannot solely rely on manual intervention; human supervision remains essential but must be supported by rapid detection, containment, and response mechanisms.
Original title: Primera notificación de una brecha de datos personales causada por un ataque ejecutado mediante un agente de IA
AEPD blog post
- Who this binds
- This AEPD blog post provides advisory guidance to all organisations processing personal data within Spain, under the General Data Protection Regulation (GDPR).
- What changed
- The AEPD has highlighted that risk analyses for data processing must now explicitly incorporate attacks assisted or executed by artificial intelligence, moving beyond generic references to malware or unauthorised access.
- What to check
- Organisations should review their current data processing risk analyses to ensure explicit inclusion of AI assisted or executed attacks and update their incident response plans to account for machine speed threats.
- What this does not mean
- This does not mean that all AI models or their providers are inherently compromised or designed for malicious activities, as the AEPD explicitly states that the use of a specific AI model does not imply this.
Send this to your team
The Spanish Data Protection Agency (AEPD) has received the first notification of a personal data breach executed by an AI agent, indicating that organisations must now explicitly incorporate AI assisted attacks into their data processing risk analyses and review incident response times.