Content ID seems pretty robust* for non-live content. I uploaded a copy of one of the few wide-release U.S. movies in the public domain ("His Girl Friday" [0]) and was hit with a takedown notice instantaneously as the upload finished. I didn't contest it at the time, and I wonder if it's because Content ID was triggered by the Columbia Pictures fanfare/logo in the opening.

[0] https://archive.org/details/his_girl_friday

edit: By "robust", I actually mean, "a bit overly proactive", in that the system seems to be efficient in its detection, but the system's design seems unavoidably vulnerable to false positives.