I think the best answer is, nobody knows. The previous equilibrium for content scraping for search engines on the internet was already at times an uncomfortable one. But I agree that from a game theory perspective, "the AI bots take and give nothing back in return" is not just hyperbole, it's the actual situation. If Google is successful in what seems to be its plans and it becomes a box where you type a question and Google gives you an answer and only a vanishing fraction of the users click through to any underlying website, that instantly eliminates the entire value proposition for vast swathes of the web to actually be on the web.

Something has to happen or Google will end up starved and locked out of everything, by means both technical and legal. Then nobody gets anything.

I don't have the answer as to what happens next, and I doubt anyone else who proclaims one super confidently. But we can do some constraints analysis. There is no world where everyone works for free so Google and other AI engines can get all the value from the content, so we can eliminate those possibilities. I think we can safely discard the world(s) in which all content production just stops. However, off the top of my head, it's hard to get much tighter than that, and that definitely leaves a world where effectively everything everywhere ends up going pay-to-access.

Microtransactions have, to date, failed comprehensively, though, so the constraints on what "everything is pay-to-access" gets weird without them.

And there is never guarantee that there is any solution to any set of constraints. Things can end up overconstrained in reality as easily as a math problem. I don't actually think it'll go that way, but when analyzing this question I think it's important to not let "but $SOMETHING just has to have some way to work, because... uh... it has to!" Let the constraints do the talking. You could end up with a scenario where all content of any value is locked down, and it's fundamentally difficult and expensive to ever access or discover it, and consequently the entire content production industry radically contracts compared to its current size, if there is no pragmatic solution to microtransactions that is low-enough friction to get over the psychological and economic hurdles that have killed it to date. If everything is locked behind "macrotransactions" that's a much smaller commercial web. Probably a much higher quality one, too, but at a pretty stiff cost.