EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

AI at Risk in the EU: It’s Not Regulation, It’s Implementation

Judith Arnal

European Journal of Risk Regulation · 2025-09 · 16 (3) · 1168-1177

DOI: 10.1017/err.2025.19 ↗

Kurz gefasst

Fachartikel in der Zeitschrift European Journal of Risk Regulation, online erschienen am 27.03.2025. Verfasst von Judith Arnal. Die Arbeit wurde bisher 4-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The implementation of the General Data Protection Regulation (GDPR) in the EU, rather than the regulation itself, is holding back technological innovation. The EU’s data protection governance architecture is complex, leading to contradictory interpretations among Member States. This situation is prompting companies of all kinds to halt the deployment of transformative projects in the EU. The case of Meta is paradigmatic: both the UK and the EU broadly have the same regulation (GDPR), but the UK swiftly determined that Meta could train its generative AI model using first-party public data under the legal basis of legitimate interest, while in the EU, the European Data Protection Board (EDPB) took months to issue an Opinion that national authorities must still interpret and implement individually, leading to legal uncertainty. Similarly, the case of Deepseek has demonstrated how some national data protection authorities, such as the Italian Garante, have moved to ban the AI model outright, while others have opted for investigations. This fragmented enforcement landscape exacerbates regulatory uncertainty and hampers EU’s competitiveness, particularly for startups, which lack the resources to navigate an unpredictable compliance framework. For the EU to remain competitive in the global AI race, strengthening the EDPB’s role is essential.

Zitieren

APA

Arnal, J. (2025). AI at Risk in the EU: It’s Not Regulation, It’s Implementation. European Journal of Risk Regulation, 16(3), 1168-1177. https://doi.org/10.1017/err.2025.19

Vancouver

Arnal J. AI at Risk in the EU: It’s Not Regulation, It’s Implementation. European Journal of Risk Regulation. 2025;16(3):1168-1177. doi:10.1017/err.2025.19

BibTeX
@article{Arnal2025,
  author = {Arnal, Judith},
  title = {AI at Risk in the EU: It’s Not Regulation, It’s Implementation},
  journal = {European Journal of Risk Regulation},
  year = {2025},
  volume = {16},
  number = {3},
  pages = {1168--1177},
  doi = {10.1017/err.2025.19}
}
Kostenlos starten →