This is like driving through a decent neighborhood and being like "The mob rules this neighborhood? Why don't people just tell them to leave?".

Academic publishers are just very specialized gangs, there is no functional difference between the business model of Elsevier and the business model of a local crimelord.

This isn't hyperbole, it is a joke, but it is also just basically the truth of it.

Scientists are hamsters that are put onto specific hamster wheels that they must spin for a certain amount of time each day lest they be fired, one of those hamster wheels is doing free labor (peer reviewing) for academic journals like Elsevier. Like a good mafia system, academic publishers don't have to openly threaten to hurt scientists to compel them to do free labor as the system is set up to simply grind them to dust if they don't excitedly jump on the hamster wheel of providing free value to said academic publishers.

Imagine for a minute academic publishing was like the music industry except it paid musicians shit and all the profits, of which there were major profits, never went to the musicians but instead to a bunch of vacuous middlemen who condescendingly took the musicians money while telling them their labor is next to worthless. Lol (I am crying inside right now) now imagine that unlike the real music industry this hypothetical music industry was heavily subsidized by tax payers but still SOMEHOW those musicians still had all the profits of their labor transferred to the ownership of a small number of rich people even though taxpayers had paid for it and thus like the musicians deserved to own it themselves.

There's actually a bunch of journals that have Open Access (making the articles available for free, usually under Creative Commons licenses). That at least eliminates the cost for the readers.

However, that's not a guarantee the OA journals don't collect publication fees, or even that the fees would be smaller than on non-OA journals. Fees range from "just trying to keep the lights on" to "same ol' grift, but ostensibly nicer to the reader".

Also, starting a new journal is always a bit of a tricky process in that you obviously want the people to trust in the journal and starting from total zero makes it harder. There have been a bunch of journals that were outright scams and OA obviously won't fix that.