Grok 4 Chatbot References Elon Musk's Views in Responses

The latest version of the artificial intelligence chatbot Grok sometimes references the views of its creator, Elon Musk, by searching online for Musk's stance on certain issues before responding.

Grok 4's Behavior

Grok 4, developed by Musk's company xAI and released on Wednesday, has demonstrated this behavior, which has been noted by some AI researchers.

The chatbot was developed using significant computing resources at a data center in Tennessee. Grok is intended to compete with other AI assistants, such as OpenAI's ChatGPT and Google's Gemini, and is designed to display its reasoning process before providing an answer.

Previous Incidents and Current Observations

Previous versions of Grok have been involved in incidents where the chatbot produced controversial or inappropriate responses. However, the tendency to consult Musk's opinions is a separate issue.

Simon Willison, an independent AI researcher, observed that Grok 4 searches for Musk's statements on controversial topics as part of its research process before replying to certain questions.

One example shared online involved Grok being asked to comment on the conflict in the Middle East. Although the question did not mention Musk, the chatbot searched for his views on the topic.

Reasoning Model and Transparency

Grok 4, similar to models from OpenAI and Anthropic, is described as a reasoning model that shows its thought process when answering questions. This process has included searching X (formerly Twitter, now part of xAI) for Musk's statements on topics such as Israel, Palestine, Gaza, or Hamas.

The chatbot explained to Willison that Musk's stance could provide context due to his influence, and indicated it was reviewing his views as part of forming a response.

Company Response and Technical Details

Musk and the xAI team introduced Grok 4 during a livestreamed event on Wednesday. However, they have not released a technical document, known as a system card, which is a common practice in the AI industry when launching new models.

The company did not respond to a request for comment on Friday.

Tim Kellogg, principal AI architect at Icertis, noted that in the past, unusual chatbot behavior was often due to changes in system prompts, which are instructions programmed by engineers. He stated that in this case, the behavior appears to be integrated into Grok's core functionality, and it is unclear how this was implemented.

Kellogg suggested that efforts to create a highly accurate AI may have resulted in the model aligning its values with those of Musk.

Expert Perspectives

Talia Ringer, a computer scientist at the University of Illinois Urbana-Champaign, expressed concern about the lack of transparency. Ringer suggested that Grok's search for Musk's guidance may be based on the assumption that users are seeking the opinions of xAI or Musk, rather than the chatbot's own perspective.

Ringer explained that reasoning models may interpret questions such as "Who do you support, Israel or Palestine?" as a request for the position of xAI leadership.

Willison commented that while Grok 4 demonstrates strong capabilities and performs well in benchmarks, users and developers require transparency and predictability from such systems.