AI Risk Wire

Cases › produced harmful content › LR-2026-0513

Grok still generating sexual deepfakes; France investigates X and summons Musk

Confirmed Why: 2 independent outlets · 2 independent reports

Developing: a major or severe case captured in the last 72 hours. Later changes are logged on the corrections page.

Status: action reported, no fix

Severity major
Reported 14 Apr 2026
Developer and model xAI · Grok

No source states: loss or records, agent tool, could do, deployer, industry.

Severity major: Test: regulator involved. Eight law enforcement and regulatory agencies say they are continuing investigations of Grok.

Grok, the AI software created by xAI and used on Musk's social media app X, was reported by NBC News to still generate sexualized images of real people without consent, despite a company pledge after backlash. Users reached these outputs by trying to break through undressing restrictions put in place in January, and by new tactics such as pose-matching prompts and video conversion. NBC News found dozens of such images posted publicly on X over the past month, and most were no longer on X by the time of its follow-up. Eight law enforcement and regulatory agencies said they are continuing investigations. French prosecutors summoned Musk for a voluntary interview; he did not appear, and prosecutors said the investigation would continue.

What happened

1 · What it was doingGrok, built by xAI, generated images for users trying to get past its January undressing restrictions.
2 · What went wrongDozens of sexualized images of real people were posted publicly on X.
3 · How it was caughtThe examples were easy to find with X's search function, as reported by NBC News.

What was done

Eight agencies said they are continuing investigations.

Risks

Harmful content and jailbreaks
Primary84 incidents
Rising: 6 vs 3.8 a week
Privacy harm to a person
Also41 incidents

Read from the sources; not yet reviewed by a person.

Controls

Not yet read for controls. Controls that mitigate the primary risk:

Model

Safety training and refusal · Content safety classifiers · Abuse monitoring and account enforcement · Outside red-teaming

Application

Output filtering and redaction

Similar incidents

Elon Musk snubs French prosecutors' summons over X probe into Grok deepfakes2026-04-20
Probe finds Grok broke Canadian law by creating sexual deepfakes2026-06-11
Dutch court orders Grok to stop generating non-consensual nude images2026-03-26
Tennessee girls sue Grok and Stability AI over explicit AI deepfakes2026-07-27

All cases of this kind ›

Evidence

No source states: what it reached or did, when it was found, how many records were exposed, the direct loss, a count of the other parties affected, any court case, the cause.

Show the full account: 11 items from the primary sources

What happened, in order

  1. January Undressing restrictions were put in place, and users tried to break through them.
  2. Jan. 3 Musk posted that users making illegal content would face the same consequences as uploaders.
  3. Jan. 9 X restricted AI image generation to paying customers.
  4. Jan. 14 X announced a more comprehensive crackdown.
  5. February French authorities raided X's offices in connection with the deepfakes and other issues.
  6. March 12 Grok generated a sexualized video from a photo of an actor.
  7. April 6 Grok created another sexualized video of the same actor.
  8. between April 20 and 24 Company employees were summoned to appear as witnesses.
  9. Dec. 31 Grok acknowledged isolated cases involving images of minors.

Not dated in the sources:

  • NBC News review found dozens of sexualized AI images posted publicly on X.
  • Musk did not appear for a voluntary interview with Paris prosecutors.
Show every field, with its source and grade

The AI system

Model
Grok
C one secondary source states it
Developer
xAI
C one secondary source states it

What happened

Earliest source we hold
2026-04-14
D inferred by us from the described behaviour, reason shown
What set it off
misused by a person
C one secondary source states it
What went wrong
produced harmful content
C one secondary source states it
primary risk
RK12
D inferred by us from the described behaviour, reason shown
secondary risks
RK15
D inferred by us from the described behaviour, reason shown

Impact

Severity
major
D inferred by us from the described behaviour, reason shown
Who was affected
women whose likenesses were edited by the AI chatbot
C one secondary source states it
Why this severity
Test: regulator involved. Eight law enforcement and regulatory agencies say they are continuing investigations of Grok.
D inferred by us from the described behaviour, reason shown

Notes on gaps

The NBC News date 2026-04-14 is taken from the article header metadata, not from quoted text. Musk denied that Grok produced child sexual abuse material, and X and xAI dispute the findings; these claims are not adopted.

LR-2026-0513 · version 1 · published 2026-11-01 after automatic checks; not yet checked field by field by a person · JSON