In order to make that assumption you have to first assume that they know qualitatively what is better and what is worse, that they have the appropriate skills or opportunity necessary to choose to opt in or opt out, and that they are making their decision on what tools to use based on which one is better or worse.

I don't think you can make any of those assumptions. In fact I think you can assume the opposite.

Isn't that what you yourself are doing, right now?

The average person does not choose their tools based on what is the most effective at producing the correct truth but instead on which one is the most usable, user friendly, convenient, generally accepted, and relatively inexpensive.

Yes, because people have more than one single criterion for determining whether a tool is "better."

If there was a machine that would always give me a thorough well-researched answer to any question I put to it, but it did so by tattooing the answer onto my face with a rusty nail, I think I would not use that machine. I would prefer to use a different machine even if its answers were not as well-researched.

But I wasn't trying to present an argument for which is "better" in the first place, I should note. I'm just pointing out that AI isn't going to "go away." A huge number of people want to use AI. You may not personally want to, and that's fine, but other people do and that's also fine.