Meta paused training its AI (LLMs) on EU/EEA Facebook and Instagram public data after the Irish DPC asked it to delay (June 2024)

The problem

In May 2024 Meta began notifying Facebook and Instagram users across the EU/EEA that, from a start date press dated to “June 26”, it would use their public content to train its generative-AI systems on an opt-out basis (source). Before that date the plan met organized privacy pushback: the campaign group noyb had filed “11 complaints” across EU countries, and the U.K. “Information Commissioner“‘s Office had also asked Meta to pause (source).

What happened

On 14 June 2024 Meta paused the plan. RTÉ News reported firsthand that Meta agreed to “pause plans to use personal data to train artificial intelligence” after “concerns were raised by the Irish Data Protection Commission” (source). TechCrunch independently reported that Meta agreed to “pause plans to start training its AI systems using data from its users in the European Union” (source).

The regulator (critical claim)

Meta’s lead EU regulator, the Irish Data Protection Commission, said the same day that it “welcomes the decision by Meta to pause its plans to train its large language model” using “public content shared by adults on Facebook and Instagram across the EU/EEA” (source). The DPC characterised the outcome as following “intensive engagement between the DPC and Meta” (source).

Meta’s framing (critical claim)

On its own newsroom, Meta said it was “disappointed by the request from the Irish Data Protection Commission” — “our lead regulator, on behalf of the European DPAs” — to “delay training our large language models” using public content shared by adults on Facebook and Instagram (source). The scope of the paused activity — training Meta’s LLMs on adults’ public Facebook and Instagram content across the EU/EEA — is carried both by Meta’s own statement and, independently, by the DPC and by both newsrooms (source).

A note on the weakest link

This is a reversal story, not a metric story: no numeric outcome is the headline, so no figure carries the page. The load-bearing statements come from strong sources — Meta’s own newsroom (Tier-1 first-party, an against-interest “disappointed” statement) and the Irish DPC (Tier-1 independent regulator) — and the pause is carried firsthand by two independent Tier-2 newsrooms, TechCrunch and RTÉ (source). The weakest, purely-contextual elements are the “June 26” planned start and the noyb “11 complaints” count: these come from the TechCrunch report, not from a primary filing or a Meta/DPC statement, and are left as press-attributed context to be confirmed on record, not asserted as adjudicated fact (source).

How this was verified

Method: four Wayback captures dated 2024-06-14 anchor the record — Meta Newsroom (about.fb.com), the Irish DPC (dataprotection.ie), TechCrunch and RTÉ News. The sha1-b32 of each raw Wayback payload equals its CDX digest (Meta K5YLUWGNG2Q6XJXAZGYTQIC7TXYQMHI3; DPC QVRQ74KS6BSBVEXVFT7S5WDGDGUFAJJ6; TechCrunch 5S4ZTVFNNZE6IGP2CJQIVMDBHZCBEJYQ; RTÉ FJ4NCLIJ3JPG66NQCUKCGNCWO3RASC6M). The Meta and DPC captures are stored gzip-encoded and tied on the raw payload; TechCrunch and RTÉ are plain HTML. Every quoted fragment above was byte-verified with grep -F against the specific capture it is attributed to (12/12 bind). The Meta capture used is the post-update 14 June snapshot — the earlier 10 June snapshot predates the “disappointed / delay” statement and does not carry it. Corroboration is two_independent on both critical claims: the pause-after-DPC-request (c1) and the scope-plus-framing (c2) are each carried by TechCrunch and RTÉ, not only by Meta and the DPC. Verified 2026-08-08. Conflation guard: this is scoped to the EU/EEA pause announced 14 June 2024 — distinct from Meta’s separate July-2024 Brazil (ANPD) suspension, its April-2025 EU resumption after the EDPB’s December-2024 opinion, and the LinkedIn/ICO-UK matter (a different company). Green is human-only: a recorded confirmation from Meta communications and/or the DPC press office of the exact scope, dates and status of the paused EU/EEA training is the human step.

Related case files

• X (Twitter) suspended EU user-data training for Grok after the same Irish DPC took High Court action — the same regulator, the same year, the same fight over training AI on Europeans’ public posts
• Brazil’s ANPD halted Meta’s generative-AI training with a 50k-daily-fine preventive measure — the same company’s training paused by a different regulator, the conflation-sibling this record is scoped against
• LinkedIn suspended UK generative-AI training on member data after ICO concerns — another platform pausing genAI training on users’ data under regulator pressure

Sources

1. Tier 1 — first party. Meta Newsroom · “Building AI Technology for Europeans in a Transparent and Responsible Way” · 2024-06-14 · https://about.fb.com/news/2024/06/building-ai-technology-for-europeans-in-a-transparent-and-responsible-way/
2. Tier 1 — independent regulator. Irish Data Protection Commission · “DPC’s engagement with Meta AI” · 2024-06-14 · https://www.dataprotection.ie/en/news-media/latest-news/dpcs-engagement-meta-ai
3. Tier 2. TechCrunch (Paul Sawers) · “Meta pauses plans to train AI using European users’ data, bowing to regulatory pressure” · 2024-06-14 · https://techcrunch.com/2024/06/14/meta-pauses-plans-to-train-ai-using-european-users-data-bowing-to-regulatory-pressure/
4. Tier 2. RTÉ News · “Meta pauses AI plans after data protection concerns” · 2024-06-14 · https://www.rte.ie/news/2024/0614/1454797-meta-ai/