23 ms·

I appreciate the attempt, but honestly, I'm not really sure how you would build something like this. You'd have to be able to filter audio, video, images and te

by bgdam 8y ago

I appreciate the attempt, but honestly, I'm not really sure how you would build something like this. You'd have to be able to filter audio, video, images and text. I'm not even sure if there exists a corpus of copyrighted content anywhere to compare against. You're probably going to have to build it yourself. And keep it up to date with not just newly generated content, but also content whose copyrights expired. Frankly, even for companies of the FAANG scale, this is a tough challenge, and I don't think a startup is going to be able to handle this.

Max-20 8y ago
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.