The European Commission has confirmed that it spent the past four weeks dealing with what its spokesperson called an “AI-generated cyber campaign” against its internal systems. Staff lost access for days to Sysper, the human resources platform they use for payslips, leave and timesheets, and other applications were taken offline as a precaution. The Commission says it has found no evidence of a data breach affecting its staff as a whole, but the investigation is still open and the key question, whether the attackers reached the personal data held in those HR systems, has not been answered publicly.
TL;DR
- The Commission says the incident was detected on 22 October 2026 and confirmed on 26 October, and that a publicly available AI tool was misused to carry it out.
- Politico reported that CERT-EU identified the tool as ARTEX and that Sysper, the MyPMO medical reimbursement system and translation services were disrupted for more than 10 days.
- The Commission says it has informed CERT-EU and the European Data Protection Supervisor, has seen no ransom demand and has not yet found evidence of a breach affecting staff as a whole.
What the Commission has confirmed
The story broke on Thursday, 19 November, when Politico’s Brussels Playbook reported a “huge” cyberattack on the Commission. Hours later, spokesperson Thomas Regnier took questions at the midday press briefing, and the official recording and transcript give the clearest on-the-record account so far.
According to Regnier, the incident was detected on 22 October and confirmed after an initial investigation on 26 October. The Commission then opened a full investigation and, “as a protective and preventive measure,” temporarily took down several applications, most of which have since been restored. He said the Commission is in close contact with CERT-EU, the cybersecurity service for EU institutions, and with the European Data Protection Supervisor (EDPS).
On the method, he was unusually direct. Asked whether AI was involved, he said the investigation had confirmed “this was an AI generated cyber campaign,” describing it as “a misuse of an AI tool” that is publicly available. He did not name the tool. He also described the event as “a campaign, so there was a series of incidents,” rather than a single intrusion.
Several points were left open on purpose:
- Attribution: Regnier declined to say who was behind the attack or what they wanted.
- Status: He would not say whether the attack is over, only that “many elements have been contained” and the investigation continues.
- Scale: He refused to describe the scale while the investigation runs.
- Ransom: He said he was not aware of any ransom demand.
- Personal data: “At this stage of the investigation, we have not detected any evidence of a data breach affecting Commission staff as a whole.”
That last sentence needs careful reading. “As a whole” leaves room for a breach that affects some staff, and a Commission spokesperson separately told Recorded Future News that individual systems were targeted and that “those concerned” had been informed. The honest summary is that the Commission has not confirmed a personal data breach and has not ruled one out.
What has been reported, but not confirmed by the Commission
Most of the operational detail comes from Politico, citing unnamed officials and an internal email. According to its report:
- CERT-EU detected earlier this month that the attackers had used ARTEX, a publicly available AI-powered tool for finding vulnerabilities in IT systems. An EU official told Politico that “some actors misuse such tools and have targeted the Commission.”
- Sysper and MyPMO, the medical reimbursement system, were restored earlier this week, although Sysper briefly crashed again on 18 November. Translation services were disrupted on 17 November.
- Salaries were largely paid on time, but some officials could not submit timesheets or request leave, and three officials complained about poor internal communication.
- The Corporate Management Board, the Commission’s top internal crisis body, held an urgent meeting on the “ongoing cyber incidents” earlier this month, and Commissioner Henna Virkkunen briefed the College of Commissioners on 18 November.
Euronews added that the Joint Sickness Insurance Scheme (JSIS) system was still unavailable on Thursday, according to two officials, and noted that systems like Sysper hold names, pay grades, home addresses and contact details. That is exactly the kind of data that could fuel targeted phishing against individual officials.
A correction to some early summaries: the Commission itself has not named ARTEX. It has confirmed only that a publicly available AI tool was misused. The ARTEX link comes from Politico’s reporting on CERT-EU’s findings. It is plausible, since the FBI warned on 13 November that ARTEX is being used by criminal and state-linked groups worldwide, but it should be treated as reported rather than officially confirmed.
A third incident in one year
This is not the Commission’s first security problem of 2026. In January, its central mobile device management infrastructure was hit, potentially exposing staff names and phone numbers. In March, attackers compromised the cloud environment hosting Europa.eu websites, an incident we covered as a data breach in the Commission’s AWS cloud account. Officials have not linked the three incidents.
The criticism this time is mostly about disclosure. Dutch MEP Bart Groothuis told Euronews that the Commission keeps information about intrusions internal instead of telling the public. Regnier rejected that at the podium: publicising defensive measures can undermine them, he said, and the required bodies were informed. “There was no hiding whatsoever.”
The rules that apply to an EU institution
EU institutions are not covered by the GDPR. Their data protection obligations come from Regulation (EU) 2018/1725, often called the EUDPR, which mirrors the GDPR’s breach rules and is supervised by the EDPS. Cybersecurity incident reporting is governed by Regulation (EU) 2023/2841, which set up the Interinstitutional Cybersecurity Board and gave CERT-EU its current mandate. Read together, they set out the clock the Commission is working against:
- Notify the EDPS within 72 hours. Under Article 34 of the EUDPR, the controller must notify a personal data breach to the EDPS without undue delay and, where feasible, within 72 hours of becoming aware of it, unless the breach is unlikely to result in a risk to individuals. Information can be provided in phases.
- Tell the people affected if the risk is high. Article 35 requires the controller to communicate a breach to data subjects without undue delay when it is likely to result in a high risk to their rights and freedoms, in clear and plain language.
- Warn CERT-EU within 24 hours of a significant incident. Article 21 of Regulation 2023/2841 requires an early warning within 24 hours, an incident notification within 72 hours with an initial assessment, and a final report no later than one month after that notification.
- Document everything. The EUDPR requires the controller to document every personal data breach, including its effects and the remedial action taken, so the EDPS can check compliance.
The Commission says CERT-EU and the EDPS were informed, which fits this framework, but it has not said when, or whether it treated the event as a personal data breach. The EDPS has its own investigative powers over EU bodies; it recently used them in a long fact-finding exercise into Europol’s systems. Whether it opens anything similar here will depend on what the investigation shows about the HR data.
Why HR platforms are a prize
An HR system is a dense store of personal information: identity data, home addresses, salary, family details, sick leave and sometimes health-related reimbursement records. For a public institution, it is also a map of who does what, which makes it useful for espionage and social engineering as well as fraud. Recent incidents elsewhere show the pattern, from the leak of more than 41,000 DHS employee records through a single sign-on system to a vendor incident involving PwC employee learning records.
For organisations watching this case, a few practical points stand out:
- Treat HR and payroll as high-risk systems. Apply the same access controls, logging and monitoring you would apply to customer data.
- Assume automated scanning. Tools that use AI to find and chain vulnerabilities shorten the gap between exposure and exploitation. Internet-facing HR portals need fast patching and regular testing.
- Plan for “contained but unclear.” Most investigations pass through a long phase where systems are restored but the data question is unresolved. Decide in advance how you will notify regulators in phases and when you will tell staff.
- Communicate internally. The complaints reported by Politico were less about payslips and more about silence. Employees whose data may be exposed should not learn about it from the press.
The Commission has also been dealing with other pressure on its security agenda, including a planned review of the supermarket fridge cyberattacks at its Security College. This incident puts its own house at the centre of that conversation.
Frequently asked questions
Was staff personal data stolen?
It has not been confirmed either way. The Commission says it has found no evidence of a data breach affecting staff as a whole, while the investigation continues and individual systems that were targeted are still being examined.
Did the Commission say ARTEX was used?
No. The Commission confirmed the misuse of a publicly available AI tool but did not name it. The ARTEX identification comes from Politico, which said CERT-EU detected its use.
Who has to be told about a breach at an EU institution?
The European Data Protection Supervisor, within 72 hours where feasible, unless the breach is unlikely to pose a risk to people. Affected individuals must be told without undue delay if the risk to them is high. Significant cyber incidents also go to CERT-EU on a 24-hour, 72-hour and one-month timeline.
Why does the GDPR not apply?
The GDPR covers companies and national public bodies. EU institutions, bodies and agencies follow Regulation 2018/1725, which applies essentially the same principles and gives the EDPS the supervisory role.
The test is still ahead
The Commission has contained much of the damage and restored most services, and it says it followed its reporting procedures. The harder test is still ahead: an evidence-based answer, delivered to staff first, on whether the data in its HR systems left the building. How it handles that answer will say more about its security culture than the attack itself.