Ireland's DPC opens an investigation into Google's GDPR compliance over using the personal information of EU and EEA users for training foundation AI models
Google's lead privacy regulator in the European Union has opened an investigation into whether or not it has complied …
Context & Ripple Effects
Ireland’s privacy regulator has already made Google’s handling of personal data a recurring enforcement focus, including an earlier inquiry into its advertising-data collection and a separate location-data inquiry. The AI question had also surfaced when the DPC required Google to delay Bard’s EU launch over insufficient privacy information, making this a move from product-readiness scrutiny to the data used beneath foundation models: the delayed Bard launch
First-order effects
- Google must answer the DPC’s questions about its legal basis and safeguards for using EU and EEA personal information in foundation-model training, putting that processing under formal GDPR review.
- EU and EEA users’ data practices become the immediate subject of regulatory scrutiny, rather than only the behavior of a finished AI product.
Second-order effects
- Other AI developers using regional user data for training face a clearer enforcement risk, particularly after the DPC’s contemporaneous action over X’s training of AI systems with EU user data.
- Google may need to devote more compliance work to documenting training-data provenance, notices, and user-data controls before expanding AI features in the region.
Third-order effects
- If regulators sustain this approach, foundation-model training will become a distinct GDPR enforcement surface, alongside long-running oversight of advertising and location-data processing.
- The practical boundary between data that is available to a platform and data it may repurpose for model training is likely to be set increasingly through regulator decisions rather than company policy alone.
The trend: European privacy enforcement is extending from how platforms use personal data in services to how they reuse it to build foundation models.