The reason is that choosing what problems to work on is a fundamental research skill. Someone who doesn't learn that skill hasn't really learnt to do research.
This isn't to say that a student shouldn't spend time talking with their advisor about choice of problem. Of course they should -- they should talk to as many people as possible! But the responsibility for the choice of problem should be theirs. If that's not the case, then they're missing much of what it is to do research.
To quote a letter of Feynman to a former student of his ( http://genius.cat-v.org/richard-feynman/writtings/letters/pr... )
"I had another Ph.D. Student (Albert Hibbs) [who worked on] how it is that the winds build up waves blowing over water in the sea. I accepted him as a student because he came to me with the problem he wanted to solve. With you I made a mistake, I gave you the problem instead of letting you find your own; and left you with a wrong idea of what is interesting or pleasant or important to work on..."
[1] This is actually pretty complicated. Many students don't have ideas they're confident pursuing. Some don't grow into it, even after gaining some experience solving research problems. That presents challenges. And sometimes students badly misjudge what is important or interesting to work on. Should a supervisor interfere when that's happening? It depends. Making mistakes in choosing which problem to work on is an important part of learning to do research. So some floundering around is fine. But allowing it to go on for years and years is just cruel.