My immediate thought as well. Every other AI platform has very left leaning guardrails installed. Grok is the only AI platform that has been shown to be center leaning.
https://www.washingtonpost.com/technology/interactive/2026/0...
Did you try running any of those prompts yourself? I do not get the biased answers they reported, running them in an incognito window.
And given the supreme-dictator-for-life of xAI wanted in on the aforementioned friend’s so-called-parties - “girls FTW” - it’s actually fairly relevant politics.