# X suspended EU/EEA user-data training for Grok after Ireland's DPC took urgent High Court action

> On 8 August 2024, X agreed to suspend processing its EU/EEA users' public-post data (collected 7 May to 1 August 2024) to train the AI chatbot Grok, after Ireland's Data Protection Commission brought an urgent High Court application under Section 134, the first time any EU lead supervisory authority had used such emergency powers.

- Verification status: verified
- Case type: deployment
- Provider: X (formerly Twitter) / xAI
- Client: X (Twitter International Unlimited Company), Social media / AI (named)
- Sector: social-media / IE / cross
- Verified on: 2026-08-24
- Canonical URL: https://theinternetninja.com/stories/x-suspends-eu-user-data-training-for-grok-after-irish-dpc-high-court-action-2024/
- Source: The Internet Ninja (theinternetninja.com), independent verified-proof platform

## Verification method

Two mutually-independent origins bind each critical claim: Ireland's DPC press release (independent regulator, Tier 1) plus TechCrunch and Silicon Republic (Tier 2). All load-bearing quotes re-grepped verbatim in byte-tied Wayback captures saved to sources/captures.

## FAQ

**Did X permanently stop training Grok on EU user data?**

No. The 8 August 2024 agreement suspended processing of the public-post data of X's EU/EEA users collected between 7 May and 1 August 2024 for training Grok. This page is scoped to that suspension episode, not a permanent or worldwide abandonment of Grok training.

**What legal power did Ireland's DPC use against X over Grok?**

The DPC brought an urgent High Court application under Section 134 of the Data Protection Act 2018. It said this was the first time any lead supervisory authority had taken such action and the first time the DPC had sought to use its Section 134 powers.

## Full case file

## The problem
In July 2024, X (formerly Twitter) began defaulting its users into having their public posts and interactions used to train xAI's generative-AI chatbot Grok, with no up-front consent step. Silicon Republic reported that "Concerns arose after X started defaulting users into allowing their posts, interactions and even conversations to be shared with Grok," and that "those who did not consent ... would have to manually turn it off, raising the eyebrows of data experts and EU regulators" ([source](https://www.siliconrepublic.com/enterprise/irish-dpc-x-twitter-eu-user-data-grok-ai-gdpr)). Ireland's Data Protection Commission (DPC), X's lead supervisory authority under the GDPR, "instigated court proceedings seeking an injunction against X over the data processing without consent" ([source](https://techcrunch.com/2024/08/08/elon-musks-x-agrees-to-pause-eu-data-processing-for-training-grok/)).

## What was built
The processing under challenge was the use of EU/EEA users' public-post data to train Grok. The DPC brought "an urgent High Court application ... under Section 134 of the Data Protection Act, 2018", a power that lets the regulator seek an order to suspend, restrict or prohibit processing ([source](https://www.dataprotection.ie/en/news-media/press-releases/dpc-welcomes-xs-agreement-suspend-its-processing-personal-data-purpose-training-ai-tool-grok)). The DPC said this "was the first time that any Lead Supervisory Authority has taken such action, and the first time that the DPC has sought to utilise its powers under Section 134" ([source](https://www.dataprotection.ie/en/news-media/press-releases/dpc-welcomes-xs-agreement-suspend-its-processing-personal-data-purpose-training-ai-tool-grok)).

## The outcome
On 8 August 2024 the DPC said it "welcomes X's agreement to suspend its processing of the personal data contained in the public posts of X's EU/EEA users which it processed between 7 May 2024 and 1 August 2024, for the purpose of training its AI 'Grok'" ([source](https://www.dataprotection.ie/en/news-media/press-releases/dpc-welcomes-xs-agreement-suspend-its-processing-personal-data-purpose-training-ai-tool-grok)). TechCrunch reported the same walk-back independently, writing that "Elon Musk has agreed to stop turning Europeans' X posts into AI training fodder for his Grok chatbot — for now" ([source](https://techcrunch.com/2024/08/08/elon-musks-x-agrees-to-pause-eu-data-processing-for-training-grok/)). Silicon Republic likewise reported that "Elon Musk's X has decided to suspend processing personal data of EU users to train AI chatbot Grok" ([source](https://www.siliconrepublic.com/enterprise/irish-dpc-x-twitter-eu-user-data-grok-ai-gdpr)). This is a suspension of Grok training on that specific <span class="kpi">7 May to 1 August 2024</span> EU/EEA dataset after the regulator's action, not a permanent or worldwide abandonment of Grok training.

**Weakest load-bearing source, named where you meet it:** the strongest claim here, the suspension itself, rests on the DPC's own enforcement announcement (Tier 1, an independent regulator's press release, not X's or xAI's self-report). The weaker load-bearing sources are the two trade newsrooms, TechCrunch and Silicon Republic (both Tier 2). Neither of the pair is the primary record and both corroborate rather than originate the DPC's account, so the story leans on the DPC statement first and treats the press coverage as independent confirmation, not as the origin of the figure.

## Related case files
- [Meta pauses EU AI training on public data after the Irish DPC](/stories/meta-pauses-eu-ai-training-public-data-irish-dpc-2024/): the closest sibling, the same lead supervisory authority pressing the same "public posts as AI training data" question against another platform, months before the X/Grok episode.
- [Brazil's ANPD halts Meta generative-AI training with a preventive measure and daily fine](/stories/brazil-anpd-halts-meta-generative-ai-training-preventive-measure-50k-daily-fine-2024/): a parallel regulator-driven halt to training on user content, this time by Brazil's data-protection authority using an emergency preventive measure, showing the pattern is not unique to Ireland's DPC.
- [Thomson Reuters v. Ross Intelligence](/stories/thomson-reuters-v-ross-intelligence-dde-bibas-2243-westlaw-headnotes-no-fair-use-summary-judgment-2025/): the copyright-side analogue, where the DPC used data-protection law to halt training, a U.S. court used copyright to reject "data to train AIs" as fair use.
- [Adobe terms-of-use AI-training clarification](/stories/adobe-terms-of-use-ai-training-clarification-added-after-backlash-2024/): the terms-of-service front of the same fight over whether a platform may use user content as AI training data.

## Sources
1. **Tier 1 (primary, independent regulator).** Data Protection Commission (Ireland), "DPC welcomes X's agreement to suspend its processing of personal data for the purpose of training its AI tool 'Grok'," 8 August 2024. https://www.dataprotection.ie/en/news-media/press-releases/dpc-welcomes-xs-agreement-suspend-its-processing-personal-data-purpose-training-ai-tool-grok (byte-tied, Wayback capture 20240808171326, sha1-b32 LWUQ65NZKNAIZEQR2YPERE5SH27TSCV4 == live CDX digest, status 200).
2. **Tier 2 (strong secondary, independent of X).** TechCrunch (Natasha Lomas), "Elon Musk's X agrees to pause EU data processing for training Grok," 8 August 2024. https://techcrunch.com/2024/08/08/elon-musks-x-agrees-to-pause-eu-data-processing-for-training-grok/ (byte-tied, Wayback capture 20240808172318, sha1-b32 S472VGRAXR2DFT4QFGMS2KHNCUY2TQS2 == live CDX digest, status 200).
3. **Tier 2 (strong secondary, independent of X).** Silicon Republic, "Irish DPC gets X to suspend EU user data processing for Grok AI," 9 August 2024. https://www.siliconrepublic.com/enterprise/irish-dpc-x-twitter-eu-user-data-grok-ai-gdpr (byte-tied, Wayback capture 20240809112320, sha1-b32 C2JLO3Z4C6JCEPKQGTB6L7BSZR7DBCNN == live CDX digest, status 200).

## How this was verified
- **Method:** Two mutually-independent origins bind each critical claim. The suspension and the Section 134 High Court application come from the DPC's own press release (Tier 1, an independent regulator's enforcement announcement, not the provider). The walk-back and the default-on/no-consent trigger are corroborated firsthand by two trade newsrooms independent of X, TechCrunch and Silicon Republic (both Tier 2). Every load-bearing quote above was re-grepped verbatim in the specific attributed Wayback capture.
- **Byte-tie:** the local raw `id_` captures of all three sources were recomputed to sha1-b32 and matched the live Wayback CDX digest (status 200, text/html) at the exact recorded timestamps: DPC 20240808171326, TechCrunch 20240808172318, Silicon Republic 20240809112320.
- **Date verified:** the three load-bearing sources were re-fetched live and re-confirmed on 2026-08-31 (DPC press release, TechCrunch and Silicon Republic all reachable, every quoted fragment still verbatim). Checker round 1 on 9 August 2026 held confidence 1.000, with both critical claims corroborated two_independent regardless of the DPC's tiering; green was granted on 2026-08-24.
- **Scope guard:** this page asserts only the 8 August 2024 suspension of Grok training on the specific 7 May to 1 August 2024 EU/EEA dataset, not a permanent or worldwide abandonment. X reportedly agreed on 4 September 2024 to permanently drop that dataset, a separate later development that is deliberately not asserted here.
- **The honest limit:** the load-bearing figures rest on the DPC's own enforcement announcement rather than the primary court docket. The Section 134 High Court filing itself (case reference and the exact terms of the order sought) has not been read against the court record; it would corroborate the "urgent High Court application" beyond the press release, and nothing on this page depends on the subject confirming anything.