The vast majority of science papers are created with a solution in mind that's looking for a problem. Someone pays for an intended outcome and they work their way towards it.
Related:
The scientists fighting the epidemic of fraudulent science research
This is strikingly similar to the downfall of Quora that's on the frontpage right now and I agree with the answers on that discussion thread. The solution, to me, is to establish more moderation. Which is to say, incentivize reviewers from all over the world to read and critique any scientific work instead of publishers hoarding all of the profit by themselves.
But then you'd just get fake reviews, and networks of fake actors boosting each others fake articles.
And it's not gonna get better with the "help" of AI