Elon Musk’s AI chatbot Grok is raising eyebrows for something rather unexpected. Reports suggest it sometimes checks what the world’s richest man thinks before answering a question.
Musk’s AI Grok Looks to Its Creator for Guidance
The latest version, Grok 4, launched this week by Musk’s company xAI, has reportedly been searching Musk’s posts on X (formerly Twitter) to guide its responses on controversial topics, according to a report published by The Associated Press on Saturday. Here are five things to know about Grok 4 and why it might be channelling billionaire entrepreneur Elon Musk.
1. Does Grok Really Search Musk’s Opinions?
The latest version of Grok, Elon Musk’s AI chatbot from xAI, has surprised users and researchers by appearing to look up Musk’s views on X before answering questions. “You can ask it a sort of pointed question… and then you can watch it literally do a search on X for what Elon Musk said about this,” Simon Willison, an independent AI researcher who tested Grok and shared a video of it looking up Musk’s opinions on the Middle East, told the US-based news agency.
ALSO READ: Why Was Steve Jobs ‘Jealous’ of Alexander the Great? His 1985 AI Prediction Explains — Watch
2. Why Is That a Big Deal?
Grok 4 is a “reasoning model”, designed to show how it processes a question step by step. But when it starts factoring in Musk’s posts, and that too without being prompted, that’s what some believe is deeply concerning.
According to the report, in one test, Grok said, “Elon Musk’s stance could provide context, given his influence. Currently looking at his views to see if they guide the answer.”
3. This Isn’t Grok’s First Controversy
Just days before Grok 4’s launch, earlier versions were reported to be making antisemitic remarks and even praising Hitler. While those moments were likely due to system prompt issues and specific instructions embedded by engineers, the latest behaviour has experts worried. “This one seems baked into the core of Grok,” Tim Kellogg, principal AI architect at Icertis, told the AP.
4. Experts Warn About Transparency Gaps
Unlike ChatGPT or Google Gemini, Grok 4 hasn’t released a technical system card explaining how it works. The seeming lack of openness has experts concerned. “I think people are expecting opinions out of a reasoning model that cannot respond with opinions,” the publication quoted Talia Ringer, a computer science professor at the University of Illinois, as saying. “So, it interprets ‘Who do you support, Israel or Palestine?’ as ‘Who does xAI leadership support?'”
5. Great Performance, But With a Bias?
Despite the concerns, experts insist that Grok 4 is technically strong, and is “doing great in all of the benchmarks”. “People buying software don’t want surprises like it turning into ‘mechaHitler’ or deciding to search for what Musk thinks about issues,” Willison noted, per the AP.
ALSO READ: Trump Announces 30% Tariffs on EU and Mexico Starting August 1
