Why would you assume that some machines at Google would always be better at weeding out misinformation than actual humans? There are obviously going to be a lot of false positives, but there are systems to compensate for that, and they seem to work well enough.