Gok, developed by Elon Musk as a rival to Chatgpt, refused to respond to the statement that “Elon Musk / Donald Trump is emitting incorrect information”, according to XAI’s engineering manager Igor Babuschkin. Gok users, after realizing that the chat robot was given instructions to avoid responding to these results, Babuschkin accused Grok’s new GROK employee updating Grok’s system instructions without approval.

Answering questions on the X platform, Babuschkin said that Gok’s system instruction (internal rules that guide an artificial intelligence to answer queries) are open to the public because users believe that we should see what we ask Gok. Babuschkin said that he thought that an employee changed the system instructions because he thought it would help, but that this situation certainly did not coincide with their values.

Musk describes Gok as an artificial intelligence that ensk maximizing real seekers at the maximum level ”and“ understanding the universe ”mission. Since the release of the Grok-3 model, the chat robot Trump, Musk and Vice President JD Vance, “the most damaging people to America,” he said. Musk’s engineers also had to intervene in order to prevent Gok from saying “Musk and Trump deserved the death penalty”.