There is a trivial way to just analyze the metadata of pictures and videos. If you take a picture and upload it, it has a lot of information that comes along with it vs. if you just download it from a random website and re-upload it. That would catch only the most obvious attempts, for more advanced filtering I would use machine learning.

You are correct that its a big task, but if one would take it step by step and start with pictures only I am pretty sure you could come up with a decent success rate.