Technology

Meta begins using European data to train AI with GDPR safeguards

Sunday, 20 April 2025 12:02AM UTC

After a year-long delay to comply with EU privacy rules, Meta will incorporate Facebook, Instagram, and WhatsApp European data into AI training. Users retain the right to object under GDPR, with a clear process established to protect personal information.

Meta has announced that it will begin using data from its platforms in Europe to train its artificial intelligence (AI) models, following a twelve-month delay to comply with European Union privacy regulations. This development means that interactions on Facebook, Instagram, and WhatsApp will be incorporated into the training of generative AI systems, such as those designed to create text, images, or automated responses.

The company, owned by Mark Zuckerberg, has been using data for AI training purposes in the United States already, but the European rollout reflects efforts to align with the European Union’s General Data Protection Regulation (GDPR), which provides specific rights to users regarding the handling of their personal data.

According to Meta’s announcement, the types of data that could be used include publicly shared posts, photos with captions, comments, profile information, and other publicly available content. Notably, the company excludes the contents of WhatsApp chats and data from users under 18 years of age from this process.

A key feature of GDPR is the right to object. Users residing within the European Union, the European Economic Area (EEA), or Switzerland are entitled to submit an objection to prevent Meta from using their personal data for AI training. This extends even to data that is publicly accessible, as such content can still contain personal information protected under the law.

Meta has created a dedicated process to enable users to exercise this right. Individuals can visit the Meta Assistance Center online, accessible via www.facebook.com/help/contact/, where they must complete a form indicating their country of residence (to confirm GDPR applicability), full name, and the email address linked to their Meta account. To prevent fraudulent submissions, Meta requires identity verification either through email or a verification code.

Once the objection is submitted, Meta is obliged to comply with GDPR, responding typically within 30 days, although processing times may vary. If the objection is accepted, the company will cease using the individual’s personal data to train its AI models.

If users do not submit an objection, Meta is permitted to continue including their public content in AI training datasets. The company states that it anonymises data; however, the potential for identifying individuals cannot be entirely eliminated.

Besides the formal objection, the article suggests that users take practical steps to limit data exposure, such as adjusting privacy settings to restrict post visibility to friends or specific groups, refraining from posting sensitive information like children’s names or addresses, routinely checking privacy configurations on Facebook and Instagram, and considering the deletion of older public posts that may already have been incorporated into AI datasets.

The process of objecting to the use of personal data for AI training is described as straightforward and free of charge, underscoring the importance of digital rights management within the evolving landscape of AI development in Europe.

Source: Noah Wire Services

More on this

https://about.fb.com/news/2025/04/making-ai-work-harder-for-europeans/ - This page from Meta announces the start of AI training using interactions from European users, confirming the company’s plan to include data from Facebook, Instagram, and WhatsApp in AI model training after a delay to comply with EU regulations.
https://thehackernews.com/2025/04/meta-resumes-eu-ai-training-using.html - This article details Meta’s resumption of AI training in Europe with adult public data only, excluding WhatsApp chats and accounts of users under 18, and explains the opt-out option for EU users under GDPR.
https://www.france24.com/en/live-news/20250414-meta-to-train-ai-models-on-european-users-public-data - France24 confirms Meta’s use of public data and user interactions on its platforms, including Facebook and Instagram, to train AI models in the EU, emphasizing the incorporation of interactions with Meta’s AI.
https://techcrunch.com/2025/04/14/meta-to-start-training-its-ai-models-on-public-content-in-the-eu/ - TechCrunch covers Meta’s announcement that AI training on public content in the EU will proceed, noting the previous pause due to GDPR concerns and the availability of user objection under EU data protection law.
https://www.facebook.com/help/contact/ - This is the official Meta Assistance Center contact page where European users can submit their objection to Meta’s use of their personal data for AI training, as described in the article.

Noah Fact Check Pro

The draft above was created using the information available at the time the story first emerged. We’ve since applied our fact-checking process to the final narrative, based on the criteria listed below. The results are intended to help you assess the credibility of the piece and highlight any areas that may warrant further investigation.

Freshness check

Score: 8

Notes: The narrative discusses recent developments regarding Meta's compliance with GDPR for AI training data use, indicating a current rollout after a 12-month delay. There are no outdated references such as past roles or deceased individuals. The topic is very topical given ongoing AI and data privacy debates in 2024-2025, suggesting it is fresh news.

Quotes check

Score: 9

Notes: The narrative does not contain direct attributed quotes but summarises Meta’s announcement and GDPR provisions accurately. The lack of direct quotes reduces opportunities to verify earliest usage, but the reported information aligns with publicly known policies from Meta and GDPR. Thus, it is likely original reporting or a direct summary from official Meta statements.

Source reliability

Score: 7

Notes: While no specific publication name is given, the detailed and accurate description of GDPR rights and Meta's AI training policies suggests a reliance on credible official information, probably from Meta’s own announcements or trusted technology news. However, without explicit attribution to a major reputable news outlet, some uncertainty remains.

Plausability check

Score: 9

Notes: The information is plausible and consistent with ongoing regulatory and technological trends in AI development and data privacy in Europe. The description of GDPR rights, objection mechanisms, and Meta’s approach matches established legal frameworks and industry practices as of 2024. No extraordinary or unverifiable claims are made.

Overall assessment

Verdict (FAIL, OPEN, PASS): PASS

Confidence (LOW, MEDIUM, HIGH): HIGH

Summary: The narrative reliably reports current Meta policy on using European platform data for AI training after GDPR compliance delays. It aligns with known regulations and corporate actions, contains no outdated or unverifiable claims, and appears to be based on original or official announcements. Despite missing direct attribution to a major media outlet, the level of detail and legal accuracy supports a high confidence rating.

Meta
Artificial intelligence
GDPR
Privacy
Data protection