Even abusive crawlers and scrapers are acting as agents of real humans, just as your browser is acting as your agent. I don't even know how you could reliably draw a reasonable line in the sand between the two without putting some group of people on the wrong side of the line.
I suppose the ultimate solution would be browsers and operating systems and hardware manufacturers co-operating to implement some system that somehow cryptographically signs HTTP requests which attests that it was triggered by an actual, physical interaction with a computing device by a human.
Though you don't have to think for very long to come up with all kinds of collateral damage that would cause and how bad actors could circumvent it anyway.
All in all, this whole issue seems more like a legal problem than a technical one.
Or the AI people could just stop being abusive jerks. That's an even easier solution.
While that is probably good advice in general, the earlier commenter wanted even the abusive jerks to have access to his content.
He just doesn't want tools humans use to access content to be used in association with his content.
What he failed to realize is that if you eliminate the tools, the human cannot access the content anyway. They don't have the proper biological interfaces. Had he realized that, he'd have come to notice that simply turning off his server fully satisfies the constraints.
That would be easier. Too bad it won't ever happen.