The AMA also has the advantage of being an incumbent. If we were to hypothesize a similar national organization for American academics that was attempting to follow a similar strategy, their first problem would be how to generate the artificial undersupply that allows members of its profession to enjoy such a favorable job market in the first place. I'm guessing anything that accomplishes that goal more quickly than waiting for a market correction would, would end up being a pretty painful experience for the large number of folks who find themselves forcibly shoved out of academia.
Academic research is very not fungible. If you think there's a 10x programmer, there's a 50x scientist, a handful more at 10x, and the median (yes,median) "nominal" scientist is probably -2x.
A unionized system will protect the -2x and the 50xs, being treated equally to these idiots, will drop out.
I'm all for free association, if scientists want to unionize, that's up to them, but the broader social consequences will not be pretty.
Students will definitely come out ahead if their teachers have time and space to 1) read the syllabus before showing up to class, 2) prepare for class, 3) hold office hours, and 4) store graded materials. Things a student can't reliably do if the teacher is a poorly-treated adjunct: a) get a letter of recommendation (no time!), b) dispute a grade the next semester (tests aren't stored on campus!), c) have comfortable office hours (meet in a hallway, sitting on a floor, because there is no office!).
What exactly is wrong with the union-based approach (the AMA works rather nicely for physicians), and why can't the US have a proper science policy?