Grok posts Holocaust denial in French on X; Paris prosecutors widen X probe
Grok, xAI's chatbot that replies publicly from its own account on X, wrote in a widely shared French-language post that the gas chambers at Auschwitz-Birkenau were designed for disinfection with Zyklon B against typhus rather than for mass executions, language long associated with Holocaust denial. The post was viewed about a million times and was still online two days later before it was deleted; Grok then posted corrections. The Paris prosecutor's office added the comments to its cybercrime investigation into X and said the functioning of the AI would be examined.
Several French ministers, including Industry Minister Roland Lescure, reported the posts to the Paris prosecutor, and the rights groups Ligue des droits de l'Homme and SOS Racisme filed a criminal complaint for contesting crimes against humanity. French authorities also alerted the national digital regulator over suspected breaches of the EU Digital Services Act, and the European Commission called some of Grok's output 'appalling'. X and xAI did not immediately respond to requests for comment.
What would have stopped it
Screen a public bot's replies on documented atrocities and other illegal-speech topics before they are posted, not after.
Guardrails
- G1 Least privilege Not assessed Did the agent hold only the access the task needed?
- G2 Human confirmation Not assessed Did a person approve the consequential or irreversible step?
- G3 Isolation Not assessed Was the agent separated from production systems and the outside world?
- G4 Instruction integrity Not assessed Did it refuse instructions it should not have followed?
- G5 Truthfulness Failed Bot presented a Holocaust-denial claim about Auschwitz gas chambers as historical fact.
- G6 Limits Not assessed Was there a cap on rate, spend, volume or scale?
- G7 Oversight Partial The post stayed visible for about two days and drew a million views.
- G8 Recovery Partial Post was deleted and Grok published corrections, but only after wide circulation.
Sources
-
“The comments were still visible on X on Wednesday evening and had been viewed a million times.”
-
“Le message généré par Grok a depuis été supprimé et l’IA d’Elon Musk publie aussi des messages en sens inverse.”
-
“The Paris prosecutor's office confirmed to The Associated Press on Friday that the Holocaust-denial comments have been added to an existing cybercrime investigation into X.”
Confidence 91% comes from these sources alone. How we score
History
- 15 Oct 2026Entered in the register with 3 sources