Poate UE să reglementeze împotriva deepfakes?
Instrumentele de inteligență artificială precum Grok de la xAI, SORA 2 de la OpenAI și Nano Banana Pro de la Google permit utilizatorilor de internet să creeze deepfake-uri – imagini și videoclipuri care înfățișează persoane făcând lucruri pe care nu le-au făcut niciodată – cu mai multă ușurință ca niciodată. Deoarece devine din ce în ce mai dificil să se facă diferența între imaginile generate de IA și fotografiile și videoclipurile reale, poate UE să implementeze măsuri de protecție împotriva deepfake-urilor?
spinner.loading
spinner.loading
Mere days into the new year, users on X began using the platform’s in-built artificial intelligence, Grok, to create sexually explicit images, asking the AI to take the clothes off photos posted by women on the platform without their consent.
This is just one example of “deepfakes”, which can now be easily created by anyone online. Deepfakes have skyrocketed in recent years, with some eight million expected to have been shared online in 2025. As generative AI gets better at creating believable images and videos, experts and internet users alike beg the question: can the EU protect us against deepfakes?
Protected under GDPR - for now
The ChatEurope team spoke with Dr. Gianclaudio Malgieri, Associate Professor of Law & Technology at Leiden University, to understand the legal framework surrounding the use of our personal data to create deepfakes. From his perspective, the landmark General Data Protection Regulation (GDPR) already protected Europeans, in principle, to a certain extent.
“When deepfakes use someone’s face or voice, that's typically personal data,” said Malgieri. “Our face is biometric, and it's sensitive data under certain conditions.” Hence, this personal data enjoys certain protections under the GDPR.
However, this protection could soon be lifted, thanks to the European Commission’s Omnibus proposals targeted at revising the GDPR and the AI Act.
Fun fact - the AI Act isn’t even fully in force yet, with a big chunk of its regulations initially meant to take effect in August 2026. This date has now been pushed back by an entire year - which means the EU would no longer be the first to strictly regulate AI.
What’s more, the protections against the spread of AI-generated content proposed by the Commission are far from adequate, said Malgieri.
Watch the video above for the full story!
Tanishk Saha