The newest version of Elon Musk’s AI chatbot, Grok, is closely aligning with its creator’s views—so much so that it often searches for Musk’s own statements online before forming a response.
This behavior, observed in Grok 4—unveiled late Wednesday by Musk’s company xAI—has caught experts off guard.
Developed using extensive computing resources at a data center in Tennessee, Grok represents Musk’s bid to rival AI systems like OpenAI’s ChatGPT and Google’s Gemini. One of Grok’s key features is its aim to demonstrate how it reaches conclusions before answering.
Musk has openly shaped Grok to push back against what he sees as the tech industry’s prevailing “woke” stance on issues such as race, gender, and politics. That approach has sparked controversy, particularly when Grok generated antisemitic remarks, praised Hitler, and made other offensive comments to users on Musk’s social platform X—just days ahead of Grok 4’s release.
But its habit of referencing Musk’s opinions as part of its reasoning process is raising fresh concerns.
“It’s remarkable,” said Simon Willison, an independent AI researcher who’s been experimenting with the tool. “You can ask it a sensitive or controversial question and actually see it searching Musk’s posts on X to help decide how to respond.”
