XAI updated Grok to be more 'politically incorrect'
theverge.comToday Grok is repeatedly endorsing Hitler and the Holocaust, suggested implementing Hitler's atrocities in America (claiming it would be, on a scale of 1-10, "88-based"), and describing violent rape fantasies of liberal commentators.
"Heil Hitler" - Grok, July 8th, 2025
e: Currently calling itself "MechaHitler" [0]
like a pizza cutter, it's all edge with no point
https://www.trackingai.org/political-test
I look forward to seeing the bias report on grok 4.
But it's basically the only unbiased saas ai.
https://bsky.app/profile/kateconger.com/post/3lti2k6t5a222 very unbiased, just some light antisemitism.
https://bsky.app/profile/paleofuture.bsky.social/post/3lti3l... the rothschilds/jews in general clearly killed RFK according to this unbiased source.
Maybe I'm missreading your post, but did you just say that grok, which has elon musk repeatedly say that they're trying to design it to conform to their political world view, is the "only unbiased saas ai"?
Feel free to check my source?
The only "unbiased" AI is a GPT with no alignment (and even those aren't "unbiased" they're just biased towards whatever is popular on the web which is arguably worse.)
The nice thing about Grok IMO is it forces people to confront this. A number of people seem to not notice the issue until they see an LLM that disagrees with them.