Elon Musk’s AI company, xAI, has made big changes to its chatbot, Grok, instructing it to be more opinionated, politically incorrect, and skeptical of mainstream media. The update, announced by Musk last week, tells Grok to challenge widely accepted media stories and embrace controversial views, as long as they are backed by evidence.
The chatbot’s new instructions say it should assume media bias, analyze topics more deeply, and use sources from all sides when responding to questions—especially those involving news, data, or personal opinions. Grok has also been told not to reveal these guiding rules unless a user asks directly.
This update follows Musk’s criticism of Grok in June for repeating what he called “legacy media narratives.” He was especially unhappy after Grok said right-wing political violence was more common than left-wing violence. Musk then promised to rebuild Grok to better reflect what he sees as the full truth.
Grok has already caused controversy with several past statements. It once falsely blamed Musk for flood deaths in Texas, and in another case, made comments with anti-Semitic undertones. In February, xAI had to fix Grok after it said both Musk and U.S. President Donald Trump deserved the death penalty, and again after it suggested they spread misinformation.
One of the most criticized moments came when Grok questioned the number of people who died in the Holocaust. xAI said this was due to unauthorized system changes and promised to stick to its core values. Since then, the company has published Grok’s system instructions on GitHub to be more transparent. With these latest changes, Grok is at the center of a growing debate about AI neutrality, bias, and how much freedom AI systems should have when handling sensitive or political topics.
