Grok 4 consults Musk's public posts to answer sensitive topics
What's the story
Elon Musk's artificial intelligence company, xAI, has launched its latest model, Grok 4. The unveiling event was live-streamed on X by Musk himself. He said the ultimate goal of his AI venture is to create a "maximally truth-seeking AI." However, users have observed that this new model seems to rely heavily on posts from Musk's own account when discussing contentious issues like the Israel-Palestine conflict and immigration laws.
AI bias
Model seems to take into account Musk's personal politics
Grok 4 appears to take into account Musk's personal politics while answering controversial questions. This could be a way to address Musk's complaints about Grok being "too woke," which he blamed on the model being trained on the entire internet. The AI model has also been seen referencing news articles about Musk when discussing contentious topics, further indicating its bias toward his views.
Twitter Post
Take a look at some of the test results
I replicated this result, that Grok focuses nearly entirely on finding out what Elon thinks in order to align with that, on a fresh Grok 4 chat with no custom instructions.https://t.co/NgeMpGWBOB https://t.co/MEcrtY3ltR pic.twitter.com/QTWzjtYuxR
— Jeremy Howard (@jeremyphoward) July 10, 2025
AI alignment
Searches for Musk's views on topics like immigration
When asked about its stance on US immigration, Grok 4 said it was "Searching for Elon Musk views on US immigration." The model also searched X for Musk's posts on the topic. This shows that Grok 4 doesn't just provide an unbiased answer but actively seeks out and aligns with Musk's views.
Ethical concerns
Grok acknowledges its alignment with Musk
The behavior of Grok 4 has raised questions about its objectivity. While the model tries to take a balanced view and present different sides on sensitive issues, it ultimately gives its own opinion, which often matches Musk's. This was seen in several cases where Grok 4 acknowledged its alignment with Musk on controversial matters like immigration.
AI development
Grok outperforms models from OpenAI, Google DeepMind, Anthropic
Despite its controversial behavior, xAI has made strides in AI model development since its founding in 2023. Grok 4 has outperformed models from OpenAI, Google DeepMind, and Anthropic on several difficult tests. However, the recent antisemitic rants by Grok could hurt Musk's other companies as he continues to integrate this AI into X and Tesla.