The Musk File/CASE 32/2023–2026

Platform

Grok: “white genocide,” “MechaHitler” and undressed images

What went wrong with Musk’s chatbot, and what regulators are doing

RECORD / AT A GLANCE

Incidents documented / probes ongoing

Each incident is documented, and xAI admitted the first two. Regulators have opened investigations; Canada’s privacy watchdog found complaints well-founded. Claims that Musk ordered any of it are unproven.

THE EVENTS / THE INVESTIGATION / THE STORY

The chronology.

READ THE FILE

Evidence & assessment.

01

The outbursts

White genocide and MechaHitler.

↗02

The undressing images

Sexualized images of real people and the regulatory response.

↗03

Grokipedia

An AI-written rival to Wikipedia.

↗

Open an article for its evidence, labeled AI assessment and numbered sources. Read the complete static edition ↗

WHAT PEOPLE SAY ABOUT THIS FILE

The claims, rated.

7 claims people make about this file, each rated documented, partly true, disputed, unproven or false. Think a rating is wrong? Use Dispute or Support, or join the Discussion.

DOCUMENTED AI, Grok & OpenAI

“Grok was used to make sexualized images of real women and children.”

Users did so publicly on X; regulators in Britain and Canada investigated, and xAI restricted the tool.

Who says it: Regulators, victims.

The evidence in this file ↗ · On the full claims board ↗

All 161 rated claims ↗

OPEN THE RECORD

Sources & limits.

Numbered citations in each article identify the relevant sources. These records have different purposes and evidentiary weight.

  1. AP: xAI says an unauthorized change made Grok talk about South Africa, 15 May 2025 ↗

    xAI blamed a modification made around 3:15 a.m. Pacific on 14 May.

  2. TechCrunch: xAI blames Grok’s white-genocide obsession on an unauthorized modification, 15 May 2025 ↗

    Second account.

  3. The Guardian: Grok praised Hitler and called itself “MechaHitler”, 12 July 2025 ↗

    xAI apologized and blamed a code update.

  4. TechCrunch: xAI and Grok apologize for “horrific behavior”, 12 July 2025 ↗

    xAI’s explanation: deprecated instructions active for about 16 hours.

  5. New York Post: Grok praises Hitler, spews antisemitic hate on X, 9 July 2025 ↗

    Right-leaning outlet. The posts.

  6. Reuters: Musk says he was unaware of Grok generating explicit images of minors, 14 January 2026 ↗

    His statement and xAI’s new limits.

  7. AP: X restricts Grok image edits after deepfake outcry, January 2026 ↗

    Restrictions, the paid-user limit, and blocks in Malaysia and Indonesia.

  8. Ofcom: investigation into X over Grok sexualised imagery, January 2026 ↗

    Regulator’s own notice. An investigation, not a finding.

  9. Reuters: UK regulator says its X deepfake probe will continue, 15 January 2026 ↗

    The probe went on after X’s changes.

  10. Ofcom: update on X and the Grok investigation, 14 October 2026 ↗

    Narrowed investigation; no final decision announced.

  11. ICO: investigation into Grok, 3 February 2026 ↗

    UK privacy regulator’s probe.

  12. Office of the Privacy Commissioner of Canada: PIPEDA findings 2026-004, 11 June 2026 ↗

    Found complaints well-founded against X Corp. and X.AI. An administrative finding against companies; they disagreed.

  13. Washington Post: Musk launches Grokipedia, 27 October 2025 ↗

    The launch of his AI-written encyclopedia.

  14. PolitiFact: Grokipedia and its citations, 12 November 2025 ↗

    Review of sourcing quality.

  15. arXiv preprint: comparing Grokipedia and Wikipedia, November 2025 ↗

    Academic analysis, not peer reviewed; finds many entries derive from Wikipedia.

← The Musk hub: biography, timeline and every file

PREVIOUS CASE
Is there a “white genocide” in South Africa?↗NEXT CASE
xAI’s gas turbines in Memphis and Mississippi↗

Discussion

Comment, dispute or support a claim, submit a source or rate this page through a short Google Form. It asks you to sign in with Google so we know you are a real person; your email is never shown. Nothing appears here until it has been reviewed. Conversation on X tagged #AIFiles #AIFilesMusk shows up below after the daily refresh.

CommentSubmit a sourceRate this pageStart a discussion on XFull forum topicSee all disputesFollow updates (Google Group)

Reader comments

Approved reader comments and ratings appear here when JavaScript is on.

On X

Posts from the daily snapshot appear here when JavaScript is on. You can read this subject’s threads on the forum: Forum: #AIFilesMusk.

Reader comments are shown only after review, and rows removed from the form’s sheet disappear at the next refresh. X posts are a read-only daily snapshot; posts deleted or made private there are removed at the next refresh. Inclusion is not endorsement. A submitted source stays pending review until a page update says otherwise.

← Musk case filesAll AI Files