As I understand it, YouTube, even with the biggest and best content library and matching system, can't take the risk to be held liable when copyrighted content gets through the filters. Especially with Deep Learning leading the system, they likely wouldn't be able to explain how their content ID system works if the EU asks them questions about the specifics.
https://www.youtube.com/saveyourinternet/
As I understand it, YouTube, even with the biggest and best content library and matching system, can't take the risk to be held liable when copyrighted content gets through the filters. Especially with Deep Learning leading the system, they likely wouldn't be able to explain how their content ID system works if the EU asks them questions about the specifics.