A monkey hammering gibberish on a keyboard can prove new math given sufficient time. That's a low bar to set. The question is if the signal-to-noise ratio is high enough for it to be worthwhile.