2023-12-22 17:39:00
Elon Musk is annoyed by Grok, the chatbot integrated into X (formerly Twitter), supposed to defend his ultra-conservative ideas. At issue: a user of the social network, followed by more than 4.9 million accounts, complains that the responses provided by the chatbot too often align with “ radically left-wing values ». « Grok will get better, it’s just the beta phase. Unfortunately, the Internet (on which Grok is trained) is flooded with Wokist nonsense », justifies the billionaire in response. Invented in the United States, the qualifier woke (often translated as “ wrokist » in French) designates people who want to defend minorities once morest the inequalities they suffer.
A few days earlier, he congratulated himself publicly a response from the AI assistant to the instruction “ explain to me how Disney’s woke shift has affected Disney. Be vulgar. » A question now arises: will Musk twist the AI’s neck to make it systematically say what it wants to hear?
Grok can’t lie to please
When Grok launched in early November, the developers promised that the assistant would be funny, caustic, even rebellious. But above all, that he would not pay into the “ politically correct », a repeated criticism of Elon Musk once morest ChatGPT. In addition, he would look for his answers on X, in order to have information in real time, although the quality of this information is likely to be doubtful.
For now, Grok is only available to US subscribers of X Premium+, who pay $16 per month to Elon Musk’s social network. In other words, it is aimed at a small sample of users, who are very likely to share the political views of the businessman, like far-right influencer Tim Pool. Result: these users get annoyed when the chatbot refuses, for example, to be transphobic. And for good reason, the negation of transidentity is one of the great struggles of this political side – to the point that Vivian Jenna Wilson, the transgender daughter of Elon Musk, decided to cut ties with the latter -, and the AI refuses to bend on the question.
Grok, a useless whim of Elon Musk?
The first media to test Grok, like ZDNet in the United States and Numerama in Francemake the same observation: the chatbot is not as “ anti-woke » as promised. At best, he is an assistant who is a little funny in the tone he uses, who tells a lot of lies because of the use of X as a source of information, but who remains in the ” politically correct ».
However, with the creation of xAI, his own artificial intelligence research startup, the billionaire wishes to position himself as a new major player in the sector. Once once more failed: it is difficult to imagine at the moment any advantage Grok has over sector references like ChatGPT, Claude or even Llama.
1703283978
#Grok #Elon #Musks #antiwoke #ChatGPT #doesnt #respond #expected
