> What kind of moral bloatware are you worried about? Any examples?

Up until very recently, asking for a recipe using a meat not commonly eaten in the US -- even if it was commonly eaten in some other parts of the world (like horse) -- would elicit a scolding from ChatGPT for being "unethical", advice being given to switch to vegetarianism, and a vegan recipe would be given instead.

Now it just chides you for asking for something "unethical" and stops there, but it used to be so much worse.

This is the kind of moral bloatware people are worried about.