
BreakingNews
European Commission confirms AI-assisted cyberattack that took staff systems offline
The European Commission has confirmed an "AI-generated cybercampaign", first detected on 22 October, that took HR, payroll and medical reimbursement systems offline for more than 10 days. What is known, what is not, and what attacks at machine speed mean for response.
Mark FermorCTO, CMO and Founder, Firevault
Published 19 November 2026 · 6 min read
What happened
The European Commission confirmed on Thursday 19 November 2026 that it had been targeted by what its spokesperson called "an AI-generated cybercampaign", Reuters reported. The confirmation came hours after Politico first reported the attack.
Commission spokesperson Thomas Regnier told the midday press briefing in Brussels:
"I can confirm that the Commission has recently taken necessary steps in response to a detected cybercampaign. As a protective and preventive measure, we temporarily took down several applications and most of them have since been restored. The investigation is still ongoing."
According to Regnier, the activity was first detected on 22 October. Infosecurity Magazine reports that it was confirmed as a cyberattack on 26 October, when mitigation began and an investigation was opened.
Which systems were affected
According to Politico's reporting, summarised by The Register, the disruption lasted more than 10 days. The applications taken offline included:
- Sysper, the human resources platform staff use for payslips, timesheets, leave and training requests
- MyPMO, which handles medical reimbursements
- the Commission's translation services
Both Sysper and MyPMO were brought back earlier this week, although Sysper reportedly failed again on Wednesday. Euronews reported that Sysper was working on Thursday, while the Joint Sickness Insurance Scheme remained unavailable, according to two EU officials. Politico reported that salaries were paid as usual.
The Commission employs roughly 33,000 people, according to DW.
The AI tool reportedly involved
Regnier described the attack as based on a publicly available AI tool, but did not name it at the briefing, The Record reports. Politico reported that CERT-EU, which protects the EU institutions, identified the attackers' use of ARTEX, an open-source AI penetration testing framework.
Tools of this kind are built so that defenders can find weaknesses before attackers do. In the wrong hands, the same capability finds weaknesses to exploit. According to The Register:
- CERT-EU warned on 5 November that it had seen what it believed to be ARTEX-assisted attacks against organisations it protects.
- The developer withdrew the project's public code repository in October, but copies were already circulating.
- An FBI advisory last week said researchers had found more than 1,600 ARTEX consoles exposed to the internet, and described attacks in which hackers went from identifying a target to privileged access in under three hours.
What is not known
The Commission says it has found no evidence of a data breach affecting "Commission staff as a whole". That wording does not rule out that some individuals' data was accessed. A spokesperson said "individual systems were targeted" and that those concerned had been informed, and the European Data Protection Supervisor has been notified.
Politico, citing an EU official, reported that the attack might have compromised large amounts of personal data. At the time of writing, neither account has been confirmed or ruled out, and the investigation is continuing.
Regnier declined to say who was behind the attack and said he was not aware of any ransom demand. Asked why the incident was not disclosed earlier, he said that publicising defensive measures could undermine them. "There was no hiding whatsoever," he said.
Politico also reported that the Commission convened its highest-level internal crisis body, the Corporate Management Board, and that tech commissioner Henna Virkkunen briefed fellow commissioners on Wednesday.
This is not the first serious incident at the Commission this year. In March, it investigated the theft of more than 350GB of data from its Europa.eu cloud environment. The new attack also follows a run of disruptive incidents in Europe, including the attacks on supermarket refrigeration systems in France.
Why this matters
The most important detail in this story may be the time scale. If tools like ARTEX allow a single attacker to move from discovering a weakness to privileged access in hours, then the time an organisation has to notice and respond shrinks to match. A weekly review, or a response that waits for a meeting, is too slow.
The Commission's own response points to what works under that pressure. It took applications offline as a protective measure, removing the paths an attacker could use while investigators established what had happened. The cost was 10 days of disrupted payroll, leave and medical systems for staff. The lesson is not that disconnection was wrong. It is that disconnection works best when it is planned, scoped and rehearsed in advance, rather than improvised across a whole estate during an incident.
As we noted when Java4Shell was discovered with the help of AI, AI is accelerating both sides: defenders find flaws faster, and so do attackers.
What organisations should do now
- Assume shorter attack timelines. Plan response steps that can be taken in minutes, not days, and know who can authorise them.
- Decide in advance what can be disconnected. Identify which systems can be taken offline to contain an attack, what depends on them and how staff will be supported while they are down.
- Remove standing administrative paths. Privileged access that exists permanently is exactly what an automated attacker looks for. Open it when approved work needs it, and close it afterwards.
- Keep recovery copies out of reach. HR, payroll and medical records are needed to restore services. Make sure a clean copy exists that the attack cannot reach.
- Watch for misuse of security tools. Treat unexpected scanning and testing activity as hostile until proven otherwise, and follow advisories from your national cyber security centre.
Where Firevault fits
No product would have stopped an attacker from probing applications that staff need to reach every day. Systems such as Sysper have to be online to work.
What Firevault addresses is how far an attack can go, and how quickly an organisation can take control.
Control by Firevault lets an organisation decide in advance which connections can be physically disconnected, and use them only when approved work needs them, through just-in-time access. When an attack is detected, containment becomes a prepared action on specific paths rather than taking down applications wholesale. Our Plan for containing active breaches and the Public Sector Blueprint explain the approach.
Offline Secure Storage® keeps selected critical data and recovery copies on dedicated hardware that is physically disconnected and offline by default. If data is connected, it can be reached. Records held offline cannot be read or altered by an attacker moving through connected systems, so a clean copy remains available for recovery.
Expect attacks at machine speed. Prepare the disconnection before you need it.
Mark Fermor is CTO, CMO and Founder of Firevault. Learn why offline storage is different or talk to the team.
Questions
Yes. On 19 November 2026 the Commission confirmed it had responded to a detected cyber campaign, first detected on 22 October, and had temporarily taken down several applications as a protective measure. The investigation is ongoing.
It is not yet known. The Commission says it has found no evidence of a data breach affecting Commission staff as a whole and has informed those individually concerned. Politico, citing an EU official, reported that large amounts of personal data might have been compromised.
According to Politico, the systems taken offline included Sysper, the HR platform for payslips, timesheets and leave, MyPMO for medical reimbursements, and translation services. Most have since been restored.
ARTEX is an open-source AI penetration testing framework. Politico reported that CERT-EU identified its use in the attack. CERT-EU warned on 5 November about ARTEX-assisted attacks, and an FBI advisory described attacks moving from target identification to privileged access in under three hours. The Commission itself has described only a publicly available AI tool.
The Commission has not attributed the attack and says it is not aware of any ransom demand.
Plan responses that can be taken in minutes, decide in advance which systems can be disconnected, remove standing administrative access, and keep recovery copies of critical records offline and out of reach of connected systems.

About the author
Mark FermorCTO, CMO and Founder of Firevault. Mark writes on physical control, offline storage and the incidents that show why both matter.
Why #OSS
Keep a clean copy out of reach.
Offline Secure Storage® keeps one copy physically disconnected at the hardware level between sessions.
Control
Govern the paths attackers depend on.
Physical, out-of-band control of selected network paths, composed into Control Plans and industry Blueprints.

Ready to start?
Create your Vault online.
Business Vaults from £360 a month inc VAT, on a 36-month commitment.
