The European Commission confirmed on Thursday that hackers used artificial intelligence tools to attack its internal systems, forcing staff payroll and medical services offline for more than ten days. The intrusion was detected on 22 October and confirmed as a cyberattack four days later.
What went down
Among the applications taken offline were Sysper, the human resources platform staff use for payslips, timesheets and leave, and MyPMO, which handles medical reimbursements. Both were restored earlier this week, though Sysper failed again on Wednesday, and translation services were also affected.
The disruption reached the top of the institution. The Commission’s Corporate Management Board held an emergency meeting on the incidents earlier this month, and tech commissioner Henna Virkkunen briefed her fellow commissioners on Wednesday about the response and the checks now running across its systems.
The AI tool behind it
The attackers used ARTEX, a publicly available AI-powered penetration testing framework, according to sources cited by Politico, which first reported the incident. The Commission itself has confirmed only that the campaign was AI-generated and used a publicly available tool, without naming it.
ARTEX lets a single operator direct fleets of AI agents that plan, probe, exploit and move through a network with little human input after the initial goal is set. Its developer pulled the code from GitHub in October, after reports that hackers were using it, but forks were already circulating.
The warning signs were public before the Commission spoke. CERT-EU, which guards the cybersecurity of EU institutions, warned on 5 November that it had seen ARTEX-assisted attacks, and the FBI said on 13 November that the tool had been used against targets worldwide since at least August.
What is still unknown
What the attackers took, if anything, is still unknown. The Commission said its investigation has found no evidence of a data breach affecting staff as a whole, and it is not aware of any ransom demand. The European Data Protection Supervisor has been notified, and the inquiry continues.
Our Read
This is the first confirmed AI-assisted attack on a government body, and the tool was not secret or expensive. It was free, public code, and it turned a job that once needed a skilled team into one a single actor could run. That is the shift defenders have to answer.
The misuse is spreading faster than the safeguards. Government tests have already shown chatbots producing viable attack plans, and agent frameworks meant for developers are being turned against them. The Commission’s breach shows the same pattern reaching the institutions that write the rules.
Why It Matters
Ten days without working HR and medical systems is not a minor outage at the EU’s executive arm. Staff could not file leave or claim reimbursements, and the institution had to run its response at the highest level. If the Commission is this exposed, smaller public bodies face worse.
The admission matters as much as the attack. Governments rarely confirm they were hit by AI-assisted campaigns, and this confirmation sets a precedent others will be measured against. It also puts the defensive case plainly: AI must be deployed for defence as fast as criminals deploy it for attack.
Background
The FBI’s advisory counted more than 1,600 ARTEX consoles exposed on the open internet by early November, and said some attacks reached privileged access in under three hours. Commission staff were at least paid on time, and most of the affected applications are now back online.
Source: The Register report and Infosecurity Magazine report


0 Comments