A German higher regional court has rejected an emergency lawsuit seeking to stop Meta from using public Facebook and Instagram data for AI training. The ruling gives Meta a provisional green light, but it does not close the legal debate around privacy, user control, and platform data in Europe.
What Meta Plans To Use
Meta announced back in April that it would begin using publicly available information from adult users to train its Llama language models. The data described in the source includes names, profile photos, comments, and likes from Facebook and Instagram.
The same information will also support Meta AI, the company’s chatbot now integrated into WhatsApp. Meta says private content like messages and chats will not be used.
That distinction matters because the case turns on the boundary between public information and personal data. A profile photo, a public comment, or a visible like may be accessible online, but the court still treated the material as personal data after Meta removes names and other identifiers.
German users have until May 26 to object or make their profiles private. The source also notes that users can opt out later, but Christine Steffen from the consumer group that filed the lawsuit warned that removal becomes difficult once data has entered an AI training process.
All data that has been entered into the AI is difficult to retrieve.
Why The Court Sided With Meta
The court found that Meta’s approach fits within European privacy law, the GDPR. It accepted that the data remains personal even after identifiers are stripped out, but it also found that Meta’s business interest in developing AI outweighed users’ privacy concerns in this emergency case.
The judges also recognized a technical problem at the center of AI training. According to the source, the court acknowledged that fully anonymizing the data is not technically feasible. That made it necessary, in the court’s view, to process information that is de-identified but still partly personal.
The ruling leaned on a December 2024 opinion from the European Data Protection Board (EDPB). That opinion says companies can use publicly available adult data, including data that can be found through search engines, under data protection rules.
According to the court, Meta meets those standards. The company swore under oath that sensitive details such as account information, home addresses, or license plate numbers will not be included in the training data.
What Users Can Still Do
The immediate options described in the source are simple but important. German users can object before May 26 or set their Facebook and Instagram profiles to private. Those steps are presented as ways to reduce or block the use of public profile information for Meta’s AI training.
The later opt-out option remains available, but the practical value of a late objection is less clear. The consumer group’s warning is that once data has already been absorbed into an AI system, taking it back is difficult.
For users, the case highlights a practical difference between preventing data use and trying to unwind it afterward. A public post or profile detail may look easy to edit or hide on a social platform, but AI training can make the question more complicated because the data may already have influenced a model.
- Public Facebook and Instagram data from adult users is at issue.
- Private messages and chats are not included, according to Meta.
- German users have until May 26 to object or make profiles private.
- A later opt-out is possible, but removal after training may be difficult.
The Bigger Legal Question Is Not Over
The decision is provisional. It applies only to the emergency proceedings, and the source makes clear that a full trial could reach a different conclusion.
That limitation is central to understanding the ruling. The court did not issue a final answer for every possible legal challenge to Meta’s AI training plans. It decided that, at this stage, the emergency claim should not stop the company from moving forward.
The Digital Markets Act (DMA) is another unresolved part of the dispute. The source describes the DMA as a European law that limits combining personal data across multiple platforms. Unlike an EU Commission ruling in April, the German court did not find a violation here.
That does not mean the DMA question has disappeared. It means the German court, in this provisional setting, did not treat Meta’s practices as violating that law.
Why This Ruling Matters For AI Training
The case is important because it shows how courts may weigh AI development against privacy objections when the data is already public. The ruling does not say that public information stops being personal. Instead, it accepts that the data can remain personal while still being usable under certain conditions.
That balance is likely to matter for platforms, users, and consumer protection groups. Meta’s position depends on public availability, adult user data, de-identification, and promises to exclude sensitive details. The opposing concern is that public visibility on a social platform does not necessarily mean users expect their activity to train AI systems.
For now, Meta has a provisional path forward in Germany for using public Facebook and Instagram data in AI training. But the ruling leaves important limits in place: the case can still proceed, a full trial may decide differently, and the practical question of how meaningful a late opt-out can be remains unresolved.