The difference is that AI can learn on its own, the better the reference, the better the answer, combined with algorithms, robot systems and sensors, very sophisticated technology will be created.
There are some obvious things like that the Sun revolves around the Earth, and if there will be enough texts with this statement how will AI find out that this info is outdated? Next it can find that another popular idea is that the Earth revolves around the Sun, and how AI can know that this answer is incorrect also? All those bots are learning on provided by human beings information and it compares what is most mentioned, probably in some authoritative sources, but info can be not totally accurate or out of date and relying on that AI will evaluate all that correctly is IMO too thoughtlessly.
Yeah, Of course, no matter how advanced AI is, it will never be accurate. One of the most dangerous is if everyone relies on AI and there are certain groups who deliberately manipulate news or writing on the internet, that is clearly a massive mistake. Especially if people forget that AI is an aid and we need to validate it
To be honest, I only tried ChatGPT occasionally and immediately didn't like it because it would ruin my frame of mind and make me lazy to look for references to compare things.