"Oops, just a bug" only goes so far. What's interesting about content ID is that only a select group of large copyright holders have access to protecting their works through it, and we have a growing number of cases where it makes "decisions" that are not in line with copyright law.
So is it a copyright enforcement system or is it really some kind of cartel mechanism which is replacing copyright law?
Inquiring minds in the federal government might want to know.
>So is it a copyright enforcement system or is it really some kind of cartel mechanism which is replacing copyright law?
It's neither of those. Content ID is an automated detection system that Google/Youtube voluntarily put in place to help make peace with big copyright holders like Viacom to stop the lawsuits:
https://archive.is/Awz5k
Basically Youtube created a win-win deal: when our content scanner finds a pirated upload matching with your content, you can choose to share in the monetization revenue we get from ads or you can remove it.
In contrast, the DMCA copyright strike takedown mechanism is the more heavy-handed attempt at "copyright enforcement".
EDIT add to reply: >So to be more direct: Google is exhibiting non-transparent, cartel-like behavior
I can't speak to any invisible back room behavior but as one datapoint... you can be a small independent musician on a small distributor like CDBaby or DistroKid and partipate in Content ID. Example: https://support.distrokid.com/hc/en-us/articles/360013534874...
The initial motivation for Content ID fingerprinting was placating the big players like Viacom but it has since been opened up to others.
A major problem with this is that those matching uploads are not necessarily pirated uploads, but the whole system is biased to think they are.
NASA has gotten claims from news organizations who copied NASA's footage of rocket launches. Those NASA videos were not pirated (and neither where the news organization videos, but that's not the point).
Piano teachers who play public domain classic works have gotten claims from publishers like Sony.
The system is heavily biased towards he big copyright holders and against other publishers, largely regardless of the actual merits of the case.
As a sibling comment says, it's at best a win-win-loose situation.
As for the DMCA: that system, and how YouTube handles it, has severe shortcomings of its own. It's very well possible that ContentID is better than the DMCA, but that's a very small hurdle to clear.
That sounds exactly like cartel behavior. It does something to maximize the profits of a small number of Google's business partners (who in return scratch Google's back in other ways), and it claims to be a copyright enforcement mechanism, but it isn't.
Mind you when we say "it" colloquially we're all talking about content ID, but legally what we're really talking about is how Google, the firm, is making decisions.
So to be more direct: Google is exhibiting non-transparent, cartel-like behavior in selecting which of its partners to pay and which not to. It misrepresents this as copyright enforcement.
> So is it a copyright enforcement system or is it really some kind of cartel mechanism which is replacing copyright law?
Yes. In this case this is certainly because this footage is in Disney's YouTube CMS and set to be auto-claimed. The fact that we are at the whim of an admin doing the right thing and removing that content from their CMS or at the very least changing the policy attached to it is problematic at best.
Intent is significant, it can determine outcomes in court rulings. So here we have this content ID system which Google claims exists to help some Google partners manage their copyright claims. One would think that would mean it helps them determine which claims are valid and which aren't. But it didn't appear to help Disney with that in this case, rather it enabled Disney to swiftly make an invalid claim after the film entered the public domain.
So Google's tool in this case didn't help their partner manage valid copyright claims. Its just helped Disney suppress some other publisher.
We've known for many years that Steamboat Willie would become public domain in 2024. If you were really trying to help your partners manage copyright claims wouldn't you build some notion of this into the system?
Whereas if you were trying to kowtow to larger publishers in exchange for favors - maybe you would leave details like public domain out of your implementation.
So what was Google's intent here? Of course they will have arguments like "it was a bug" or "that would be too hard," but finding out if those things are true is what you do discovery for.
As far as I know, Content ID has no information about public domain content, nor does it understand fair use in any way other than by allowed match percentages that you configure (as a content owner). Content owners are under agreement to only claim content that they have the rights to. If you abuse the privilege Google gives you, which is an approval system, ostensibly Google would remove your access to use the system. (Basically no chance that would happen to Disney)
You are totally right that it should do these things, but I contend that an extralegal process that grants more rights to content owners (especially big ones) at the expense of fair use shouldn't be improved to the level of acceptability, I think it shouldn't be allowed.
> So is it a copyright enforcement system or is it really some kind of cartel mechanism which is replacing copyright law?
Kinda both, IMO.
So, talking to defence lawyers about copyright cases, even if you settle before going to court, is extremely expensive, and even if you're a well-earning software developer you really don't want to pay even the counterparty legal costs that would precede a court case (and those can count as "actual damages" from an act of copyright infringement). The copyright holder may only get part of even just those legal costs back, so they don't really want to do that either.
For both rights holders and infringers, a thing like ContentID is actually pretty good. Likewise even mere suspected (but actually innocent) infringers, as things can be taken down without having to face a huge legal bill even if you're innocent.
But the downside is the cost savings is in the form of automation, and as there is no algorithm for the truth (and only just starting to be one for "common sense") this automation can be systematically abused by people who are not rights holders. (This isn't the cartel-like behaviour, this is just straight up old fashioned fraud).
Cartels are associations of suppliers who conspire to restrict competition and inflate prices. I think copyright does this all by itself[0], but the extra part of what you wrote was "replacing" copyright law, and… kinda? To the extent that ToS are contracts, and contracts bind parties to do more than the law alone, but can't do illegal things?
But the usual implication is that a cartel is doing forbidden things, even though that's not strictly true and they can sometimes be legally required[1], and that would be something I couldn't prove — but I don't see any particular reason why the big players in video IP would even be interested in doing that given they've already got copyright law. The closest I know of is things where a studio is inspired by a YouTuber, makes a big thing, and then someone sends a take-down request to the original[2], which seems familiar from Douglas Adams' Hitchhiker's Guide[3]
[0] compulsory cartels include those forced by law[1], and there are copyright holders who lobby for the laws, in some places I'm told it's not possible to voluntarily renounce copyright (though that may have been a game of telephone involving a misunderstanding of a court ruling), and that this was behind the discussion of why there are equivalent licenses that aren't strictly "public domain": https://en.wikipedia.org/wiki/Public-domain-equivalent_licen...
[2] Pixels movie and it's antecedent, and I vaguely remember Tom Scott saying something similar as an example of the one time he used a small claims court?
> I vaguely remember Tom Scott saying something similar as an example of the one time he used a small claims court?
Tom Scott said two things about himself in YouTube's copyright system isn't broken. The world's is. [0]
1. "A TV channel from Thailand took one of my videos, played it out without permission in one of their big television shows, and then put that entire television show into Content ID. I got a Content ID hit on my original video from them, and that took a long time to sort out."
2. "The UK has [...] the Intellectual Property Enterprise Court [1][...] and that court has a small claims track. If you're an individual photographer or video maker, you can [...] issue a claim as a "litigant in person", which is the fancy British term for "without a lawyer". I've done it, because a major company ripped off one of my videos. [...] And there I was, walking in, no lawyer, thinking I had a case, thinking I had a grip on reality. Statistically, the odds weren't in my favour, but I filed successfully. And the company settled with me, they paid me to drop the claim, because they were in the wrong, and they knew they were in the wrong, and they knew I could actually get it to a judge in exchange for a bit of work and a court fee of just over £100."
He did not say these two incidents were related.
He did not say he used a court to resolve a Content ID claim.
So is it a copyright enforcement system or is it really some kind of cartel mechanism which is replacing copyright law?
Inquiring minds in the federal government might want to know.