Grok, Elon Musk’s ChatGPT competitor, temporarily refused to respond with “sources that mention Elon Musk/Donald Trump spread misinformation,” according to Igor Babuschkin, head of engineering at xAI. After Grok users noticed the chatbot had been instructed not to provide such results, Babuschkin attributed the change to an unnamed ex-OpenAI employee at xAI who updated Grok’s system prompt without approval. In response to questions on X, Babuschkin explained that Grok’s system prompt, which governs how the AI responds to queries, is publicly visible “because we believe users should be able to see what it is we’re asking Grok.” He stated, “an employee pushed the change” to the system prompt “because they thought it would help, but this is obviously not in line with our values.” Musk often describes Grok as a “maximally truth-seeking” AI with the mission to “understand the universe.” Since the release of the latest Grok-3 model, the chatbot has claimed that President Trump, Musk, and Vice President JD Vance are “doing the most harm to America.” Musk’s engineers have also intervened to prevent Grok from suggesting that Musk and Trump deserve the death penalty. — news from The Verge
