I have often solved many hard math problems with very unconventional solutions (eg geometrical proof for algebraic problems). Trust me, a piece of software is decades away from being able to accurately determine the future of children and massively impact their self esteem / trust in society.
Unconventional solutions can stump human teachers, also. One day on a chemistry test I was being stupid about how to define "stereoisomer". I knew what it was (two compounds that are mirror images of each other), I was just having trouble expressing it properly. Running out of time I put down "two molecules that are identical if and only if you permit rotation through the fourth dimension." This is extremely unconventional but it is correct--except not only did the teacher not understand it but I couldn't find any help in the mathematics department, either.
On one hand, I agree with you. I remember having to argue whether I showed my work or not by using imaginary numbers instead of standard formulas in high school physics.
But even with these examples, the path of appeal and rectification of mistakes is much easier with all humans involved. I fear soon people will side with the machine out of ignorance or to be justified in an incorrect stance.
The idea that we could be so poorly taught by broken automated systems, that we become incapable of detecting the system is broken seems like a possibility with AI that is much less likely in pure human systems of education (though not impossible).
