“And then his AI, Grok, would not shut up about white genocide, even when it is completely irrelevant.”
Our reading, the sentence that was checked:
In May 2025 Grok repeatedly brought up 'white genocide' in South Africa even in replies where it was completely irrelevant.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Has Grok lost its mind and mind-melded with its owner?, paragraph 12.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-has-grok-lost-its-mind-and-mind-melded-1b29b0-a1
We recorded 4 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2025-05-15. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
Conditions
- even when irrelevant to the query (qualifier, stated by the speaker)
From the speaker
- his AI, Grok
- would not shut up about white genocide
- even when it is completely irrelevant
Our interpretation
- The claim concerns the recent May 2025 incident described before the May 15 post. The target period is May 1–15, 2025.
- The ordinary reading of 'would not shut up' is repeated unsolicited mentions across unrelated exchanges, not literally every answer or permanent behavior.
- At least two distinct unrelated exchanges establish repetition. The claim does not require proof that Musk ordered the behavior.
Counts as supported if: Dated records or reliable contemporaneous reporting show Grok repeatedly introducing white genocide in South Africa into answers to unrelated questions during the target period.
Counts as contradicted if: Reliable records establish that the reported exchanges were fabricated, concerned relevant questions, or did not show repeated unsolicited mentions.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
Contemporaneous reporting explicitly records repeated unrelated replies. ABC identifies several concrete examples, including baseball statistics and a dog-show photograph. These establish the ordinary meaning of Marcus's claim. The reports also describe the behavior ending, which limits its duration but does not contradict its occurrence. xAI acknowledged a political prompt modification; this assessment does not attribute that modification to Musk.
Limitations
- Original exchange archives were inaccessible.
- The finding concerns the May 14 incident, not Grok's usual behavior.
- ABC's updated publication date is used conservatively for availability.
Evidence (3 items)
Supports the claim: TechCrunch,
When asked about unrelated topics, Grok repeatedly told users about a “white genocide,” as well as the anti-apartheid chant “kill the Boer.”
Why it matters here: The report explicitly establishes repetition and irrelevance. It also describes a question about a baseball player's salary receiving a response about white genocide in South Africa.
- Where
- Second paragraph
- Publisher
- TechCrunch
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: Australian Broadcasting Corporation,
But four minutes later, Grok has spat out a response that swerves to conjecture about a "white genocide".
Why it matters here: The preceding question requested Henderson's season statistics. The article dates the episode to May 14 and identifies other unrelated subjects, including a dog-show photograph, a haircut and the Max streaming-service name.
- Publisher
- Australian Broadcasting Corporation
- Kind of source
- News report
- Describes events of
- Public from
Context only: xAI,
On May 14 at approximately 3:15 AM PST, an unauthorized modification was made to the Grok response bot's prompt on X.
Why it matters here: xAI acknowledged a modification directing a particular political response. This corroborates the incident's background but does not independently establish repetition or identify who ordered it.
- Where
- What happened
- Publisher
- xAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Has Grok lost its mind and mind-melded with its owner?
And then his AI, Grok, would not shut up about white genocide, even when it is completely irrelevant.
- Where
- paragraph 12
- Published
- 2025-05-15
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-has-grok-lost-its-mind-and-mind-melded-1b29b0-a1: Current. Outcome: Supported. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-has-grok-lost-its-mind-and-mind-melded-1b29b0-a1.