AI: Usually (not always) factual, never
quite right. There's always this vague sense that it doesn't understand the context in which it's carrying on, nor its audience. Whatever "personality" it tries to adopt, if any, feels canned and is reliably off-putting.
It's no help that AI is usually sent off on a suicide mission by people who are using it because they're lazy. That intent inevitably shines through. I know immediately when I'm dealing with an AI chatbot, and immediately start trying to circumvent it. Soon, most companies will have no such means of circumvention... because actual intelligent, insightful human beings are too expensive (and don't easily scale to volume after
n mergers.)
(The last time I had to deal with a known human who, by rights, owed me money, I got an obvious AI wall of text email back. My wife happened to speak to someone else in the same position we were in, and that person had gotten the exact same email - except the sender had neglected to delete the ChatGPT prompt before sending.

)
The only thing worse than all of these complaints about AI, is the inescapable hellscape we'll have entered when it's gotten good enough to correct for them.