The European Commission today confirmed a cyberattack with artificial intelligence tools. Detected on 22 October and confirmed on 26 after initial checks, this campaign remains under investigation. His spokesman Thomas Regnier did not publicly identify his authors or his extent.
Restored applications, an open investigation
Several applications were temporarily disabled as a precaution. Most have been restored. The Commission shall maintain contact with the CERT-EU and the European Data Protection Supervisor. She doesn’t confirm that the campaign is over.
With regard to the data, the official wording is limited: at this stage, no evidence of a violation affecting Commission staff as a whole has been detected. This does not amount to excluding any leakage.
The precise role of AI remains unknown
The spokesperson confirms the use of a publicly available AI tool, without naming it. It does not specify the transactions it would have performed. He also states that he is not aware of a ransom request.
These statements therefore do not allow the attack to be described as a fully autonomous operation. The use of a tool, the allocation of a campaign and the assessment of its consequences are three different questions. Not all public records resolve them.
A separate file from the March attack
This campaign should not be confused with the incident involving the Commission’s hosting infrastructure, revealed in March. In a reply to the European Parliament dated 23 June, Executive Vice-President Henna Virkkunen indicated that the other attack was a cloud hosting service. It mentioned 39 internal clients and 18 other European entities involved.
These numbers describe the perimeter of the March file, not that of the incident detected in October. They cannot be used to encrypt the new campaign. The extent of the consequences remains to be established.
Photo archive: Berlaymont, Brussels. Jai79 / Pixabay, Libnanews fonds present since March 2019 ; date of shooting unknown. Editorial dress Libnanews.





Comments