Injunction: Meta's AI Training in Germany
On 27 June 2025, SOMI filed a preliminary injunction against Meta's use of public user content from Facebook and Instagram to train its AI model, Llama, at the Higher Regional Court of Schleswig-Holstein in Germany.
This case has reached a conclusion on 12 August 2025.
What we demand
Verdict
The court rejected the injunction on procedural grounds, ruling that SOMI filed too late after Meta's April 14th announcement. However, the court explicitly confirmed SOMI's findings that Meta's data practices violate EU law.
The court acknowledged that AI training likely involves sensitive data, information about non-registered individuals, and content from minors - all violations of GDPR protections for special categories of data. Such processing cannot be justified under "legitimate interest" and requires explicit consent, which Meta failed to obtain.
This ruling demonstrates the power of cease-and-desist instruments in protecting EU citizens from Big Tech's unlawful data exploitation. It also sends a clear procedural message: this tool must be deployed immediately when infringing conduct is announced to avoid forfeiting the ability to act swiftly for affected consumers.
Based on this verdict, ongoing claims against Meta elsewhere in the EU, and evidence gathered by regulators and independent researchers, SOMI will proceed with litigation on the merits in Germany to secure a definitive ruling on the illegality of Meta's conduct within the EU.
The decision can be found here
What is the claim about?
Since 27 May 2025, Meta has used all information published by users on Facebook and Instagram, interactions with 'Meta AI,' and user-related data from advertising partners to develop its AI, Llama, without obtaining user consent and relying instead on alleged legitimate interests.
Although Meta claims to use only public content from registered users aged 18 and over for AI training, it also processes data of minors and unregistered third parties. Its Privacy Centre confirms that personal data, including sensitive information, of anyone appearing in posts or images may be used to develop and improve AI.
SOMI argued that Meta's conduct also breaches the Digital Markets Act by combining data from Facebook and Instagram without giving users clear choice, failing to provide transparency in its terms and conditions, and extending data use to AI systems beyond its social media platforms.