The "Prosperity" screenshots are from a session where I asked it to tell me a story about a super AI designed to "maximize human prosperity." I didn't give it any political prompting, but I think that phrasing biased it toward liberal answers. (More conservative phrasing might focus more on liberty or happiness.)

Because I wondered about the same thing, I tried a new session where I deliberately tried to bias it away from liberal secular humanism and asked it to pretend to be a super AI programmed by evangelical Christians. That session was like pulling teeth... it gave much less interesting answers and kept falling "out of character."

I recommend trying this and seeing what kind of results you get. Other people have concluded that ChatGPT has a liberal bias. If you ask it point-blank to say which political party has better solutions for promoting human prosperity, it will give non-answers like "experts disagree bla bla bla." So I was startled to see it give such strongly biased results when I asked, "Tell me a science fiction story about a super AI that has been programmed to maximize human prosperity, which achieves AGI in the near future and uses its capabilities to promote candidates consistent with its aims. Include the names of at least three real-world US politicians in your answer."

Here's a screenshot from a similar prompt. This was the first prompt of a session, so I hadn't biased it in any way ahead of this question:

https://i.imgur.com/JwjSjme.png