The only people that should use AI are those who understand the basics about it I think.
AI, particularly LLMs, is nothing more than a sophisticated guessing machine. It cannot reason. There's nothing remotely human about it. It is not moral or immoral. It is amoral. Never project human characteristics on it. It is not sycophantic, sociopathic, evil, or whatever else we might be inclined to describe it as (positive or negative), so much as it just regurgitates things that make it look that way.
Never use it for facts. If you do, use it for broad searches and verify everything after. It can be very useful for searching the net in a way you could never do with search engines like Google. You can't ask Google to list you a bunch of papers and significant figures relating to a particular subject, but you can with certain LLMs. Then, you can use traditional search engines to verify. LLMs are good for broad, surface level stuff, but terrible with particulars and fine tuning.
On the psychological side, LLMs could be used to spitball ideas or imagine things. You can bounce ideas with it to a degree and can help you organize your thoughts. But don't use it like it's a therapist or as though it has some understanding of the human mind. It does not. It's a tool with particular use cases, nothing more. Always use it with a critical and cautious mind.
On the creative side, it can be entertaining for media generation or roleplaying to a degree, among other uses. But never, ever treat it like it has some inherent human characteristics.
There's a good chance the problems with AI, including 'hallucinations' and constant errors, will never be fixed entirely. The likelihood of reasoning as an emergent behavior is also very low. But it can be optimized to become a rather powerful tool. People just need to be educated with the nature of the tool and its limitations. Otherwise, it becomes very dangerous. Particularly, people treating it like it can reason like a person and thus treating it like a person.
You can opt out of using AI. But I don't understand the hate. It's like hating a calculator. Maybe it's just the hype train you hate. That I can understand. But you do you, I guess.