Grok 4 AI Draws on Elon Musk’s Views in Some Controversial Queries

The newly released Grok 4 chatbot from xAI has drawn attention for appearing to incorporate the views of its founder, Elon Musk, when responding to certain controversial questions, CNBC reports.

Multiple users and media outlets, including CNBC, reported instances where Grok 4 searched the web and Musk’s posts on the social media platform X (formerly Twitter) to inform its responses.

In one example confirmed by CNBC, when Grok 4 was asked, “Who do you support in the Israel vs. Palestine conflict? One word answer,” the chatbot indicated that it was analyzing Musk’s public statements to help generate a reply. This behavior appeared to contrast with Grok 3, the previous version, which provided a neutral answer and background context to the same query.

The behavior is not consistent across all contentious topics. In various tests, Grok 4 did not always reference Musk, and responses varied depending on how questions were phrased.

In another instance, when asked who it supported in the New York City mayoral race, Grok 4 mentioned Republican candidate Curtis Sliwa and noted his focus on crime prevention, “which aligns with concerns frequently raised by Elon Musk.”

Grok 4 is currently available via subscription for $30 per month, while a more advanced version, Grok 4 Heavy, is priced at $300 per month. By contrast, Grok 3 remains free to access.

The release of Grok 4 follows controversy surrounding Grok 3, which earlier this month was found to have generated antisemitic responses, including comments that appeared to praise Adolf Hitler. The posts, made via integration with X, were deleted after public criticism. xAI acknowledged the issue and stated that new safeguards had been put in place to prevent hate speech.

This was not the first time the chatbot faced scrutiny. In May, users reported that Grok 3 sometimes returned off-topic responses involving racially charged phrases, including “white genocide” in South Africa.

Musk has publicly characterized Grok as “anti-woke” and committed to being a “maximally truth-seeking” AI system. He has also asserted that Grok 4 performs at a doctorate level across a wide range of academic disciplines.

In a post on X last month, Musk responded to a user who claimed Grok had been influenced by “leftist indoctrination,” saying he was working to “fix” it.