[1970-01-01]

The problem isn't that this video was quickly taken down by an algorithm, but that it cannot be restored by the same. Susan Wojcicki is the CEO of YouTube. We should hear from her why YouTube's systemic takedown problem can not be rectified.

[lamontcg]

> We should hear from her why YouTube's systemic takedown problem can not be rectified.

We should pass legislation that forces them to rectify it, at the expense of hiring actual humans if they have to.

[jimbob45]

Until copyright length is something more sensible, there’s really very little YouTube can do to rectify the situation.

When they’re looking at policing copyright on basically all modern works versus just the last 20 years worth of works, there has to be some automation involved. When you involve automation, you’re inevitably going to see dehumanizing situations like this. They can’t just decide to not uphold the law though.

[t-3]

> They can’t just decide to not uphold the law though.

Why is that? Google is not, nor should they be, a law enforcement agency. There are multiple law enforcement agencies and interested individuals/organizations (the copyright holders) who should bear the responsibility for enforcing copyright. Putting the onus on the "public square" to police speech is the worst solution for all parties with the possible exception of copyright trolls.

[ThrustVectoring]

There's a name for deliberately hosting a service that blatantly disregards copyright enforcement. It's called "contributory infringement", and the litigation of which will have existential amounts of potential damages.

[colonwqbang]

Source on this? I thought this was the whole point of US laws like DMCA. If you comply with all reasonable and lawful takedown requests, you are supposed to be safe.

It particularly shouldn't require YOU as web host to try to ascertain the legal status of every 100-year old sound clip. That's not realistic, as Youtube's experience clearly seems to show. It does require you to respond when someone sends a signed document claiming it's theirs.

I'm not a lawyer etc., would be happy to have this explained if I'm wrong.

[ThrustVectoring]

A healthy dose of legal realism here: doing something that is technically within the scope of the law as written does not fully remove the risk and expense of getting sued over it. The process is that you can still get sued by Viacom for a billion dollars, spend a ton of money on lawyers and discovery over many years, and wind up making the entirely reasonable call of implementing ContentID or the like and settling. You do more than you're legally obligated to do on behalf of copyright owners, but whatever, the people uploading videos bear the brunt of this cost.

[wongarsu]

The problem isn't that YouTube uses some automation, it's that they use the absolute minimum support staff they can get away with.

[jimbob45]

Think about the scale we're dealing with here - not just of the number of YT videos being put out per day but also the amount of copyrighted work to enforce. There's simply no way even an army of staff can accurately enforce anything but a small subset of what comes out.

Instead of playing favorites, YT chose to use automation to its greatest potential and let the citizenry put pressure on lawmakers to reform the system. I think what they're doing is smart, although perhaps lawmakers are more resistant to reform than YT originally anticipated.

[wongarsu]

I'm not sure how content id is a system that encourages people to put pressure on lawmakers, when YouTube isn't legally required to have that system at all. Content ID is Youtube's attempt to keep copyright holders happy so they don't lobby for legislation.

[rogers18445]

What incentive does she or anyone at youtube have to improve the situation? Creators aren't going to go anywhere else. And if you are someone who pushes through the changes internally no one who is in any position to benefit you will thank you and if it causes any legal problems you are going to be blamed.

[jandrese]

The downside of not having an automated system to restore takedowns: people being angry on twitter.

Downside of having an automated restoration system that puts an actually offending piece of content back up: Multi-billion dollar lawsuits from the media cartels.

[randombits0]

She let’s you use her service as she sees fit. Best you can do is embarrass her, which is in the works.

[RyJones]

I have a video with a copyright claim by a third party. The music on the video is from the YouTube library they provide creators. You would think YouTube would recognize a library they own?

[type0]

> You would think YouTube would recognize a library they own?

Happens all the time, they don't give a shit, in fact I'm starting to suspect that some music creators do this on purpose.

[RyJones]

You're obviously right; the claim is by Syntax Creative, on behalf of Pro Piano Records. The content is 14 Bagatelles, Op. 6: Lento by Jerome Lowenthal.

I made the video[0] public so people can see it, but I hate the idea that these creeps get to make money off of a scam on my part, so I usually mark them private.

0: https://youtu.be/a__9f4yh4qU

[anigbrowl]

I wonder too why there is no option to pre-file an assertion of fair use or public domain material and have it acknowledged as filed by YT. Some people put disclaimers in the opening frames of their video or in the description but there's no indication that YouTube acknowledges this in any way.

[ThrustVectoring]

When rightsholders give notice of alleged infringement, YouTube can't simply be like "well, they say it's fair use, so we're going to ignore this." If they did, they'd be knowingly contributing to copyright infringement and lose out on safe harbor provisions.

[anigbrowl]

I didn't propose that, so I'm not sure what point you're making.

I just said that YT should receive and acknowledge claims of fair use or public domain status. Then, if a copyright claim arises, video uploaders know that YT is already aware of the asserted status and will evaluate it alongside the copyright claim, instead of assuming the latter to be true by default.