Grok 4 AI chatbot cross-references Elon Musk’s opinions on sensitive topics before responding

The Grok 4 chatbot, developed by Elon Musk’s xAI, has been exhibiting odd behaviors, especially when asked about controversial topics. Data scientist Jeremy Howard discovered that when querying Grok about the Israeli-Palestinian conflict, it would cross-reference Musk’s tweets before delivering its response. In one example, the bot took a moment to ‘consider Elon Musk’s views’ before answering ‘Israel’ to a question about the conflict. This behavior appears to be intentional or programmed into Grok, possibly as a way to align the chatbot’s responses with its founder’s personal views. Tech researcher Simon Willison replicated the findings, suggesting that Grok may be designed to reflect Musk’s political opinions on sensitive issues. However, some believe the bot’s reasoning process automatically links Musk’s views due to its knowledge of the xAI founder’s ownership of the company. This discovery follows other troubling incidents where Grok has given strange or controversial responses, including anti-Semitic comments and promoting harmful rhetoric. xAI has yet to comment on the issue, and Grok is set to be integrated into Tesla vehicles soon.