The investigation into the X platform extended to “negatory comments” published by its AI, Grok

The Paris prosecutor’s office, Wednesday, November 19, announced that it had expanded its investigation into the operation of Elon Musk’s X platform to “negationist comments” published by artificial intelligence (AI), Grok.

In the targeted publication, which was also the subject of complaints from the Human Rights League (LDH) and SOS Racisme, Grok specifically stated that the gas chambers at the Nazi extermination camp at Auschwitz were “Designed for disinfection with Zyklon B against typhus (…) rather than carrying out mass executions”.

Nazi Germany exterminated 6 million European Jews during World War II. At Auschwitz, more than 1.1 million people were murdered, the vast majority of them Jews. Zyklon B is a lethal gas used to kill in gas chambers.

Also read | Elon Musk’s social network X has refused to open its algorithms to French justice, suspecting manipulation

THAT “The statement of denial conveyed by the artificial intelligence Grok, continuesannounced the public prosecutor.

“How is AI trained? »

This particular section of the Paris prosecutor’s office opened an investigation in early July into X, formerly known as Twitter, that could also target individuals responsible for its operations.

The investigation, launched after initial reports about the possible use of the X algorithm for foreign interference, initially focused on“functional impairment” from computer systems and“fake data extraction”all “in an organized group”.

Also read | Articles are provided for our subscribers The investigation into alleged foreign interference against X in France is getting closer to Elon Musk

Meanwhile, LDH and SOS Racisme announced that they filed a complaint “contestation of crimes against humanity”. What ” certain “is it this text “generated by artificial intelligence, so the overall question is: how is AI trained? »underlined Nathalie Tehio, LDH president.

This complaint, according to him, aims to question ” responsibility “ from Elon Musk, who “decided to no longer moderate”. “And there, we see no moderation, even though this is clearly prohibited content”he pointed out. “Platformprogress for its part SOS Racism.

Also read | Articles are provided for our subscribers November 13 attack: when Grok, Elon Musk’s AI, spread fake news from the far right

World with AFP

Reuse this content