The Cologne Higher Regional Court has rejected an attempt to stop Meta from using public Facebook and Instagram data to develop its AI models.
A German consumer protection group tried to block Meta's plan to train AI using personal data from Facebook and Instagram, which includes profile pictures, public comments, and user interactions. The group argued Meta lacked a proper legal basis to do so under Europe's GDPR and that Meta was violating the Digital Markets Act (DMA).
The court sided with Meta, claiming:
- Meta has legitimate business interests in developing AI services tailored to different regions, and 'legitimate interest' fully qualifies as a legal basis for scraping user data.
- The court noted that Meta had informed users about how their data would be used and gave them ways to object or make their content private. The potential implication here is that if people hadn't wanted their data to be scraped, they should not have shared it publicly.
- The DMA was not violated, since using data from both Facebook and Instagram didn't violate rules against combining user data - the information was processed without identifying specific individuals.
- Regarding concerns that some sensitive data might be included in the scraping, the court said it wasn't automatically prohibited when users had made the information "obviously public."
Granted, the court's decision was not received with enthusiasm from some privacy specialists. This ruling has the potential to influence similar cases across Europe and shape how other tech companies approach AI training-related data collection.
Let's watch this space.
