Data Protection Commission (DPC) — Meta Platforms Ireland (Meta AI)
The DPC issued a statement on Meta's plan to use adult EU/EEA users' public content to train its Meta AI large language models, confirming ongoing engagement and scrutiny of the data-protection basis for the AI training.
Key takeaway — how to prevent this
AI services need a lawful basis, transparency and (for minors) age verification before launch in regulated markets — several DPAs blocked AI products first and asked questions later.
Case brief
Read from the authority's press release on 2026-09-18. Every holding carries a short verbatim quote from that text.
- Decided by
- Data Protection Commission
- Parties
- Data Protection Commission v. Meta
- Remedy
- Meta implemented updated transparency notices, easier objection forms, and dataset protections, and the Data Protection Commission required Meta to submit an evaluation report by October 2025.
- Appeal
- Appeal status not stated
Key holdings
Meta notified the Data Protection Commission in March 2024 of its intention to train its Large Language Model on adult public content from Facebook and Instagram in the EU and EEA.
“In March 2024, Meta informed the DPC of its plans to train its Large Language Model using public content shared by adults on Facebook and Instagram across the EU/EEA.”
Following an EDPB opinion and DPC recommendations, Meta reassessed its submission and set a revised start date for generative AI model training.
“Meta reassessed its proposal and provided updated documentation to the DPC in respect of its plans to begin training of its generative AI model(s) on 27 May 2025.”
The regulator required Meta to prepare a report assessing the efficacy of its safeguard measures.
“As part of our ongoing monitoring, the DPC has required Meta to compile a report”
In plain English
In March 2024, Meta informed Ireland's Data Protection Commission of plans to train its Large Language Model using public content shared by adults on Facebook and Instagram across the EU and EEA. The Data Protection Commission raised concerns about the proposed rollout, leading Meta to pause training in June 2024. The Commission then sought a formal GDPR Opinion from the European Data Protection Board in September 2024 to achieve regulatory harmonisation across Europe. Following the release of that Opinion in December 2024, Meta reassessed its proposal and submitted updated documentation regarding plans to start training its generative AI models on 27 May 2025. Following recommendations from the Commission and feedback from other supervisory authorities across Europe, Meta implemented several user protection measures. These include updated transparency notices, easier objection forms accessible in-app across Europe, longer notice periods, controls allowing users to change public posts to private, and enhanced data protection measures such as de-identification and dataset filtering. The Commission is actively monitoring the rollout of user notifications and objection forms, which apply only to public posts. Additionally, the Commission required Meta to produce an evaluation report on the efficacy of these safeguards, which is expected in October 2025.
Inclusion does not imply wrongdoing or liability. Status reflects the latest information we have — always check the official source for current status.