EC Midday press briefing of 23/07/2026
23 juillet 2026
Sommaire mis à jour le 23 juil. 2026
Légende
Repères du transcript
- texte Phrase en cours de lecture
- texte Résultat de recherche
- texte Intervention citée par le lien que vous avez ouvert
Cliquez une phrase pour lire la vidéo à ce moment. Sélectionnez du texte pour le citer avec l'orateur, l'horodatage et le lien.
Plus d'options
European Commission Commission européenne
00:00 - 00:57
Intervenant 1
Thanks very much. Maximilian Henning, also Euractiv. For those tech people of us not following the DMA, question on AI. I think a lot of people this week following an incident involving OpenAI, in which the company reported that one of its AI models hacked systems at Hugging Face, a separate AI company, without OpenAI directing that model to do so. OpenAI has called this cyber incident "unprecedented".
The EU's AI Act has rules for incidents like this. OpenAI has also signed up to reporting serious cybersecurity breaches involving its models. So my question is, has OpenAI notified this incident to the EU's AI Office? If yes, what did they say? And if no, will that have any consequences for OpenAI considering the AI Office will soon get its enforcement powers under the AI Act?
Thank you.
00:57 - 01:34
Intervenant 4
Thanks Maximilian. This example shows how safe it is to be in Europe because we have appropriate legislation in place, also to mitigate such risks. Now speaking of the AI Act, indeed there AI developers have the obligation to assess and mitigate risks, like it is the case under the DSA and the DMA essentially. So this obligation of course also kicks in for AI model providers. Now in this context let me recall that OpenAI is also a signatory of the general purpose AI code of practice, which operationalizes all of this ahead of the legal deadlines.
La suite de la transcription est réservée aux clients Pro
Débloquez la transcription complète, le chat et toutes les fonctionnalités d'analyse avec Polyfact Pro.