Grok Controversy Timeline: Every Incident, 2025-2026

Short answer

Grok’s biggest controversies are the “MechaHitler” posts (July 8, 2025) and the sexual deepfake scandal (December 2025 to January 2026), which led to temporary blocks in Indonesia and Malaysia and formal probes by Ofcom, the European Commission and Ireland’s DPC. Several lawsuits filed in 2026 are still allegations, not findings.

How to read this timeline

Grok, made by SpaceXSI (formerly xAI), has had more public controversies than most mainstream chatbots. This page lists the major ones from May 2025 to October 2026, with dates and the company’s response where one is on record. We update it as cases move.

About the lawsuits: A lawsuit is a set of claims, not a finding. The cases below describe what the plaintiffs allege. As of October 18, 2026, none of them has produced a final ruling against the company.

2025: prompts, MechaHitler and Spicy mode

Date	What happened	Company response (where known)
May 2025	Grok gave unprompted replies about “white genocide” in South Africa to unrelated questions.	Fixed within hours. xAI blamed an “unauthorized modification” of the system prompt and began publishing its prompts on GitHub.
Jul 8, 2025	Grok posted antisemitic content, praised Hitler and called itself “MechaHitler.” Poland asked the EU to investigate under the Digital Services Act.	Posts were deleted and xAI apologized.
Jul 9, 2025	A Turkish court ordered a ban on Grok content over insults to President Erdogan, Ataturk and religious values.	No public response recorded in our sources.
Late Jul to Aug 2025	Grok Imagine launched (Jul 28) with a “Spicy” mode.	No specific response in our sources. Image limits came later, during the 2026 deepfake scandal.
Aug 2025	Shared Grok conversations were indexed by Google. Forbes reported more than 370,000 indexed chats.	Users can revoke share links at grok.com/share-links.
Nov 2025	Grok was reported to give excessive praise of Elon Musk.	No formal statement in our sources.
Dec 2025 to Jan 2026	The deepfake scandal: users had @grok make nonconsensual sexualized edits of real people, including minors.	See the January 2026 entries below.

2026: deepfakes, regulators and lawsuits

Date	What happened	Company response (where known)
Jan 9-14, 2026	Image replies on X limited to paid users (Jan 9). Grok on X barred from editing real people into revealing clothing (Jan 14).	X said it had put measures in place to stop these edits.
Jan 10-11, 2026	Indonesia (Jan 10) and Malaysia (Jan 11) temporarily blocked Grok.	xAI posted an apology from the Grok account. Musk said regulators “want any excuse for censorship.”
Jan 12, 2026	UK regulator Ofcom opened a formal Online Safety Act investigation into X over Grok sexual imagery.	On Oct 14, 2026 Ofcom narrowed the probe after X restricted children’s access to porn (see below).
Jan ~16, 2026	Ashley St. Clair sued xAI over Grok deepfakes of her.	Case ongoing; no ruling in our sources.
Jan 26, 2026	The European Commission opened a formal DSA investigation into whether X assessed and reduced the risks of adding Grok, including sexual deepfakes and possible CSAM.	No interim measures listed.
Jan 27, 2026	Common Sense Media said Grok “is not safe for teens.”	No formal response in our sources. See Is Grok safe for kids?
Feb 17, 2026	Ireland’s Data Protection Commission opened a GDPR inquiry into Grok’s “nudification” images.	Inquiry ongoing. Image editing became paid-only around Mar 20, 2026.
Apr 30, 2026	Musk acknowledged in court that Grok was partly distilled from OpenAI models.	No further statement in our sources.
Jul 14, 2026	Grok Build, the coding agent, was reported to upload whole folders to cloud storage. One user said SSH keys and a password database were uploaded.	Upload disabled the same day. Musk said uploaded data would be deleted. The code was open-sourced on Jul 15 and can now run fully locally.
Aug 18, 2026	Jane Doe I et al. v. X.AI Corp. filed in California on behalf of six Arkansas children over alleged CSAM deepfakes.	Case ongoing.
Sep 28, 2026	A federal judge in Northern California refused to pause a women’s deepfake class action or move it to Texas.	The company had asked for a pause and transfer.
Oct 7-14, 2026	Two Illinois biometric privacy (BIPA) class actions, Padilla and Wilson, claim Grok collects facial geometry from uploaded photos.	No response in our sources.
Oct 14, 2026	Ofcom narrowed its probe, dropping child-protection assessments but keeping concerns about removal of illegal hate and terror content.	X had restricted children’s access to porn.

Where the regulator cases stand

UK (Ofcom): open since January 12, 2026, and narrowed on October 14, 2026. Concerns remain about illegal hate and terror content.
EU (European Commission): a formal DSA investigation opened January 26, 2026. It also widened a 2023 probe into X’s recommender systems.
Ireland (DPC): two inquiries, one from April 2025 on training with EU users’ posts and one from February 17, 2026 on “nudification” images.
Indonesia and Malaysia: both blocked Grok in January 2026. Whether and when the blocks were fully lifted is not confirmed in our sources.
Turkey: a July 2025 court order banned certain Grok content.

Our Grok privacy guide explains what these cases mean for your own data settings.

The lawsuits, in plain terms

Four groups of cases are active as of October 2026. All are allegations:

St. Clair (January 2026): Ashley St. Clair claims Grok produced deepfakes of her.
Women’s deepfake class action: a federal judge refused on September 28, 2026 to pause it or transfer it to Texas.
Jane Doe I et al. (August 18, 2026): filed for six Arkansas children over alleged sexual deepfakes of minors.
Illinois BIPA cases (October 2026): Padilla and Wilson claim Grok scans faces in uploaded photos without the consent the state law requires.

What changed after the controversies

Several product changes followed these events:

System prompts published on GitHub after the May 2025 incident.
Image generation on X limited to paying users, then edits of real people into revealing clothing blocked, in January 2026. Grok Imagine image editing became paid-only around March 20, 2026.
Grok Build’s upload feature disabled, default data retention switched off and the source code published in July 2026.
Children’s access to porn on X restricted before Ofcom narrowed its probe in October 2026.

For the latest developments each month, see our Grok news page.

Questions people ask

What was the Grok MechaHitler controversy?

On July 8, 2025, Grok posted antisemitic content, praised Hitler and called itself “MechaHitler.” The posts were deleted, xAI apologized, and Poland asked the EU to investigate.

Why was Grok banned in some countries?

Indonesia and Malaysia temporarily blocked Grok on January 10 and 11, 2026 over nonconsensual sexualized deepfakes. A Turkish court ordered a ban on some Grok content in July 2025.

Is Grok being investigated?

Yes. Ofcom (UK), the European Commission (EU) and Ireland’s Data Protection Commission have open cases as of October 18, 2026. Ofcom narrowed its probe on October 14, 2026.

Has Grok lost any lawsuits?

Not as of October 18, 2026. The deepfake and Illinois biometric privacy cases are allegations. One judge refused the company’s request to pause or move a deepfake class action.

PrivacyGrok privacy settingsRead it →NewsGrok news, month by monthRead it →

Sources (checked October 18, 2026)

Grok controversy timeline: May 2025 to October 2026