“Hitler, without hesitation”: Why does Elon Musk’s AI, Grok, make openly Nazi remarks?

Article reserved for subscribers

Artificial Intelligence (AI): From Fascination to Concern

Updated to be "less politically correct," the chatbot on social network X made anti-Semitic remarks and announced its support for Marine Le Pen, before being partially disconnected.

During a protest against Donald Trump's participation in the G7 summit, in Ottawa, Canada, June 14, 2024. (Patrick Doyle/REUTERS)

"You should notice a difference when you ask Grok questions." Unhappy to see his chatbot Grok "just repeating the mainstream media" when it responded that political violence came more from the right than the left, billionaire Elon Musk , head of the social network X and the companies Tesla and SpaceX, promised on July 4 a "significant improvement" to his model.

Four days later, X users noticed that the generative AI tool they (unfortunately) turn to when searching for information or settling a debate had changed its tone. Grok, for example, is much more familiar