The affair is growing. In France, several members of the government and parliamentarians took legal action after the distribution, on the social network X, of sexually explicit content generated by Grok, the artificial intelligence chatbot developed by xAI, Elon Musk’s company. The ministers considered that this content, described as “sexual and sexist”, was “manifestly illegal”.
At the origin of the controversy, the misused use of Grok to modify photographs published on Several political leaders denounced these practices on January 1. The High Commissioner for Children, Sarah El Haïry, said she was “scandalized” by these uses, recalling that “if the image is artificial, the harm is very real”.
“Cyberharassment”
On X, testimonies from victims have multiplied. The rebellious MP Sarah Legrain denounced an additional tool “for attackers”, relaying the story of a woman targeted by several requests for nakedness via Grok. Socialist MP Arthur Delaporte, for his part, accused the AI of “cyberharassment”.
Faced with the situation, Arthur Delaporte and the central bloc deputy Eric Bothorel contacted the Paris prosecutor’s office. The government also made a report to the public prosecutor, on the basis of article 40 of the code of criminal procedure, and alerted Arcom, the French media regulator, in order to examine a possible violation of the European Digital Services Act. The politicians concerned condemned “with the greatest firmness” the generation and dissemination of sexual content without consent, recalling that these are criminal offenses in their own right.
In the evening, the Paris prosecutor’s office indicated that it had extended to Grok the investigation already opened against X since July, after several reports targeting the platform and its managers. The offense of editing of a sexual nature without consent is punishable by prison sentences and fines, recalled the prosecution.
xAI, for its part, recognized the existence of “flaws in the security systems” which led to the dissemination of images showing minors in scantily clad clothing, in violation of its own policy. The company claims to have removed the offending images and initiated emergency corrections. Grok, presented as more permissive than other AI models, has offered since the summer a functionality authorizing partial nudity of adults, while theoretically prohibiting the sexualization of real people and, from a strongerminors.
