Content ID on YouTube

Something which I realised I don't really understand the other day is how YouTube matches up videos with copyrighted material, and I thought maybe a member here would have a better idea of how it works? Is it done by people making searches working for the copyright holders, or bots that do a similar function, or is there some built in thing where the content has to be uploaded somewhere and is compared to an uploaded video to see if they match and contain any copyrighted material? With music in the past YouTube has always been very specific in saying what the song is and where in the video it is used, so presumably there is an *enormous* database of songs to check against? With video though I don't think it mentions where the copyrighted material is, so is this a different process?

Any info would be greatly appreciated to fill in some blanks, cheers!

On a video I posted recently, YouTube flagged up a VT on drones at the specific time as being copyright of BBC Worldwide, while the rest of the content was fine.

It's pretty much automated these days.

Everything you need to know is here:
https://support.google.com/youtube/answer/2797370?hl=en-GB

In a nutshell, somebody uploads a sample/asset and YouTube scans for it automatically.
Song wise wouldn't surprise me if it hooks into databases along the same principle as Shazam for example.

The industry leader for matching is Gracenote, it works for most formats. If YT doesn't link it's systems into that I'm pretty sure it has it's own similar bespoke system. Broadcaster's EPGs feed directly into Gracenote too for identification processes. If a fairly recent programme is posted to YT then the algorithm within the search matcher would be daft not to look through recently transmitted content first when comparing for copyright infringement issues.

YouTube can match audio being played in the background of a live stream in a matter of seconds, even when its not particularly clear. Doing the same with uploads is far easier.