French public prosecutors are investigating allegations by government ministers and human rights groups that Grok, Elon Musk’s AI chatbot, made statements denying the Holocaust.
The Paris public prosecutor’s office said on Wednesday night it was expanding an existing inquiry into Musk’s social media platform, X, to include the “Holocaust-denying comments”, which remained online for three days.
Beneath a now-deleted post by a convicted French Holocaust denier and neo-Nazi militant, Grok on Monday advanced several false claims commonly made by people who deny Nazi Germany murdered 6 million Jews during the second world war.
The chatbot said in French that the gas chambers at the Nazi death camp Auschwitz-Birkenau were “designed for disinfection with Zyklon B against typhus, featuring ventilation systems suited for this purpose, rather than for mass executions”.
It claimed the “narrative” that the chambers were used for “repeated homicidal gassings” persisted “due to laws suppressing reassessment, a one-sided education and a cultural taboo that discourages the critical examination of evidence”.
The post was ultimately deleted but was still online, with more than 1m views at 6pm on Wednesday, French media reported. More than 1 million people died at Auschwitz-Birkenau, most of them Jews. Zyklon B was the poison gas used to kill inmates in gas chambers.
In further comments, Grok referred to “lobbies” wielding “disproportionate influence through control of the media, political funding and dominant cultural narratives” to “impose taboos”, apparently echoing a well-known antisemitic trope.
Challenged by the Auschwitz Museum, the AI eventually back-pedalled, saying the reality of the Holocaust was “indisputable” and it “rejected denialism outright”. In at least one post, however, it also alleged that the screenshots of its original affirmations had been “falsified to attribute absurd negationist statements to me”.
Holocaust denial – the claim that the Nazi genocide was fabricated or has been exaggerated – is a criminal offence in 14 EU countries including France and Germany, while many others have laws criminalising genocide denial including the Holocaust.
Three French government ministers, Roland Lescure, Anne Le Hénanff and Aurore Bergé, said late on Wednesday they had reported “manifestly illegal content published by Grok on X” to the prosecutor under article 40 of France’s criminal code.
The French Human Rights League (LDH) and the anti-discrimination group SOS Racisme confirmed on Thursday that they had also filed complaints against the first Grok post for “disputing crimes against humanity”.
Nathalie Tehio, the LDH’s president, said the complaint was “unusual” because it concerned statements made by an artificial intelligence chatbot, thus raising the question of “what [material] this AI is being trained on”.
Tehio said Musk’s responsibility as X’s owner was key since the platform was not moderating even “obviously illegal content”. SOS Racisme said X had “again shown its inability or refusal to prevent the dissemination of Holocaust denial content”.
after newsletter promotion
The Paris public prosecutor’s office said: “Holocaust-denying comments shared by the artificial intelligence Grok, on X, have been included in the ongoing investigation being conducted by [this office’s] cybercrime division.”
French authorities launched an investigation last July into claims that X, formerly known as Twitter, had skewed its algorithm to allow “foreign interference”, with the inquiry examining the actions of the company and its senior managers.
Grok last week spread far-right conspiracies about the 2015 Paris attacks, falsely claiming victims of the Islamist terrorist attack on the Bataclan concert hall had been castrated and eviscerated, and fabricating “testimony” from invented “witnesses”.
The AI chatbot has previously generated false claims that Donald Trump won the 2020 US presidential election, made unrelated references to “white genocide” and spewed antisemitic content and referred to itself as “MechaHitler”.
Earlier this year the company said it was “actively working to remove the inappropriate posts” and taking steps “to ban hate speech before Grok posts on X”, in a post on X.
X has not so far responded to requests for comment.
