If I may;
AI does not know the truth.
It knows only what humanity has recorded, repeated, published, believed, and fed into its systems. So if humans continue accepting man made stories as truth, AI can reproduce those stories with great confidence.
But there is another layer; AI can also learn from the question itself, the user's language, interests, profile, previous interactions, usage patterns, and the information they engage with. It can therefore become very good at determining what a person is likely to want to hear; not necessarily what is objectively true.
That distinction matters.
A convincing answer is not the same thing as the truth. A personalized answer is not the same thing as the truth. And millions of people believing the same story does not transform that story into truth.
AI can reflect humanity's knowledge, beliefs, assumptions, and conditioning. The real question is whether you are using AI to discover truth; or simply asking a machine to confirm the illusion you already believe.