Elon Musk has repeatedly stated on X that he is against censorship and in favor of free speech. However, this seemingly applies only to what he and his allies are allowed to say. According to The Verge, Musk’s chatbot Grok refused to provide responses mentioning that “Elon Musk/Donald Trump spread misinformation.”
Igor Babuschkin, chief engineer at Musk’s AI company xAI, explained that an employee updated Grok’s system prompt without approval.
The system prompt consists of internal rules that dictate how the AI responds to questions.
“An employee made the change because they thought it would help, but this is obviously not in line with our values,” Babuschkin said.
He emphasized that Grok’s system prompt is publicly available because the company believes users should be able to see what they are asking Grok to do.
Musk has previously described Grok as a “maximally truth-seeking” AI with the goal of “understanding the universe.”
Despite this, Musk’s engineers had to intervene to prevent Grok from stating that Musk and Trump deserve the death penalty.
Since the launch of the latest Grok-3 model, the chatbot has declared that President Trump, Musk, and Vice President JD Vance are the ones who “cause the most harm to America.”
These incidents raise questions about Grok’s claimed independence and Musk’s influence over the chatbot’s responses.
About The Author
