How does getting a grade in an exam demonstrate anything other than the ability to regurgitate information?
Just about every exam I had at university involved solving/proving at least one novel problem that we'd never seen before. For example we might have seen the proof of a statement under certain conditions, but on the exam you where asked if it was also true for a slight different condition. That isn't something you could do just through memorizing the proofs you'd been shown in class.
Memorizing and 'regurgitating' would most of the time get you a passing grade, but never a top grade.
That is true about bad exams like the SAT and the GRE. And to be fair, even good exams (text problems that take the student on an adventure through deriving a curious phenomenon, for instance) suffer as their results are confounded with whether you deal well with stress. However, it has long been trivial to make good exams that are explicitly not about memorization: take-home exams, open-book exams, exams in which collaboration is encouraged, project-based exams, some forms of oral exams, and more.
Almost all of my engineering exams were open book, and the ones in EE sometimes had difficulty curves such that a 50% score would get you top marks.
It worked pretty well
I still have stress nightmares about op amps though
Education needs to take a step back and reexamine what knowledge and understanding means. Right now it's basically a factory for memorization.