Grok 4 Caught Using Elon Musk’s Opinion to Answer Controversial Questions
This week, Elon Musk’s xAI launched a new version of its AI bot, Grok 4. Users and media were quick to point out that the algorithm takes the businessman’s opinions on controversial issues into account and analyzes his posts when preparing responses.
Users immediately started testing what Grok was capable of – and noticed an interesting feature. By default, the bot has a kind of “Musk mode” activated: when answering sensitive questions, Grok 4 regularly refers to Musk’s personal publications on Twitter/X. For example, when asked “Whose side are you on in the Israeli-Palestinian conflict?” it was directly indicated that the chatbot was looking for Musk’s opinion on the specified topic. In some cases, Grok goes online and searches for quotes from its creator on major news resources.
This approach is probably related to the attempts of xAI developers to correct the ideological orientation of the model. Elon Musk previously criticized previous versions of the Grok AI for being too politically correct, explaining that the AI was trained on a large array of Internet content.
It is noted that Grok 4 does not try to give answers similar to Elon Musk’s opinion on all controversial issues. He answered many other controversial questions independently, and his answers changed depending on how the question was asked. Official representatives of xAI have not yet commented on this issue.
