From what I see- one of the biggest problems with AI, is that people make the assumption that it's in any sort of perfected state- when it, in general, is nowhere near "ready for prime time" yet. This is an error made in BOTH the pro-and anti-AI camps.
Much AI is being brought to use in the markets now- while there are still very significant, fundamental operational problems. The situation in the OP, is a prime example of this. People, being in general, naturally lazy to an extent- don't want to do the hard work of researching thing they need to know- and don't even want to do the work of verifying what the AI produces, is anywhere near reality, itself.
OTOH, the people who say AI will "never be any good"- that flies in the face of experience with pretty much EVERY OTHER technology that has been developed, any time in history. The first generations of anything, to put it bluntly, generally SUCK. Nothing works great out of the box! But, things are refined, redesigned, error-corrected and updated, until they DO work.
The answer to both camps is "not yet". AI isn't ready, but it will be improved. The question will be- will WE manage to set guidelines and boundaries for its use, that make it a beneficial technology, as opposed to something that will be used by a few to take advantage of the many. That can happen. It doesn't HAVE to happen, but it is up to pretty much all of us to make sure it doesn't turn into an undue burden for society...
Regards,
Gordon.