Content ID and the Rise of the Machines

One data point: I tried submitting "His Girl Friday", one of the most famous movies to have fallen into public domain, from the copy hosted on archive.org:

https://archive.org/details/his_girl_friday

As soon as it was finished uploading, I was notified that a copyright owner (Sony Pictures, which I guess owns the remnants of the defunct studio that produced His Girl Friday) had issued a takedown notice. Amusingly, a minute later, I received the usual "Congratulations, your video is now on YouTube!" email that you get when successfully publishing a video.

After getting a takedown notice, you have the option to dispute it, which I did, providing evidence that His Girl Friday is indeed in the public domain. A short time later, I got a note saying that my dispute was rejected, and that I could choose to fight it -- but that the decision would be up to the rights holder (Sony). However, if the rights owner chose again to deny me, my YouTube account would be penalized and I could possibly lose uploading capabilities.

I didn't really care enough to fight it (though I've kept the screenshots and emails to document the process)...anyone else get to this stage and succeed (or fail)?

Content ID is possibly the worst thing about YouTube. It's stifling. It's an incessant pain for anyone who produces any type of review, analysis or critique of popular media, particularly combined with YouTube's 3 strikes system and lack of decent appeals process. Content creators frequently see their channels disappear entirely from YouTube without explanation or means of recourse, only having them restored by fans mass-emailing Google in protest.

Another good resource discussing abuse of DMCA on YouTube is Jim Sterling's rant on the subject: https://youtu.be/d9mTOq6mP2I

What the EFF article doesn't discuss deeply is that YouTube doesn't provide much assistance when a YouTuber counterclaims; which is particularly odd since YouTube has inventive in keeping their content creators happy.

Lax copyright enforcement one of the reasons Blip.tv was popular (and the site Doug Walker used) until its closing. Now, YouTube has a functional monopoly in video hosting (No, Vimeo does not count) and no reason to play nice.

What's telling is that Google / YouTube will use this technology to defend the copyrights of big media companies, but won't do anything about homemade, viral videos uploaded by individual users which get blatantly copied, re-uploaded, and monetized. If they wanted to, they could easily make it so that the "original" version of any given video was the only one users ever saw, but they don't. They know which side their bread is buttered on.

If bandwidth was symmetric, this would never have happened.

Technically YouTube and other services like it are not needed. People need a means to distribute their videos on the internet —possibly starting with a link from the web. With symmetric bandwidth and a suitable peer-to-peer protocol, there would be no need for a centralised service. Just serve your videos from your home router and be done with it.

Now that there's millions of authors instead of a single YouTube, copyright notices and subsequent takedowns suddenly becomes much harder. No more automated process, they have to (e)mail you a legal notice.

Sure, they could still go after your ISP, have it spy on your communications, and take down your very internet connection. However that would cause even more obvious human rights problems (the internet is the only medium through which freedom of speech is a reality —journals, TV, are much too guarded and centralised for the people to actually speak in them), and it would be solved just as easily as the bandwidth problem: stop the monopoly on the internet connections: we need more providers. Hundreds, thousands of them.

It's not hard either: to encourage the existence of many small providers, we just have to take down economies of scales in pricing. One solution would have the state build the lines (such as the fibre), and lease the resulting bandwidth to providers (without bulk offers).

---

That way, takedowns becomes prohibitively expensive. As it should be. Now we can concentrate on taking down the real bad guys.

The article conflates ContentID (which is not the DMCA notice/counternotice process) with the DMCA process. ContentID, AFAICT, doesn't follow the DMCA rules, and doesn't provide either side of the DMCA safe harbor -- OTOH, since its more favorable to the participants (who are the copyright holders most likely to file costly suits) it is just as good for the infringement liability side of the safe harbor, and, for the counternotice side, losing the safe harbor against users isn't important, because Google has structured there terms of service (like every other provider) to minimize the possibility that there would ever be liability to a user for taking down content or even shuttering an account, so there is no need for that side of the safe harbor.

If you're like me and are more interested in how Content ID works, check out this Google paper on the subject.

http://cs.nyu.edu/~eugenew/publications/icassp07.pdf

At a high level, the content search algorithm uses a finite state transducer which operates on trained Gaussian Mixtures.

I strongly suspect Youtube no longer wants to be a general purpose streaming site. They want to be a curated distributor, the next Netflix with a strong social component. Of course they have a paid tier now, but the next step in their transition is to encourage high-risk, low-value channels to leave.

Obviously, Youtube's claim system is intentionally designed to favor rights-holders, they would probably have half a dozen Viacom lawsuits on their hands if it weren't as aggressive as possible. But I believe the dispute resolution system is designed poorly on purpose so that channels that receive numerous takedowns will be encouraged to disappear, or to no longer use copyrighted content regardless of whether or not it's covered by fair use.

Youtube wants that content to only be distributed by the rights-holders through official or approved channels, and eventually only through the paid tier, and they want to prove to media companies that they're capable of exerting that degree of control over digital distribution. Which means they also couldn't care less about Doug Walker, Jim Sterling or GradeAUnderA. They do care about Sony and Disney and what will make money in the future.

This is just speculation on my part, but it feels true. Given their recent debacle with G+ integration, they seem to regard Youtube as a problem to be solved, rather than a service that deserves to stand on its own.

I think that Fair Use can be outscored to robots if it's done right.

Doing it right means that they should think like economists. If mass sending sending false positive DCMA requests has very small cost to the sender, it's profitable to do so and let the receiving end pay the price.

Invalid takedown request should have a cost. It could be waiting period of X days before new takedown request can be send, or it could be money needed to lift the waiting period. Google has big enough volumes that they can afford to hire few people who review cases based on statistical sampling. It should be possible to find equilibrium that minimizes the harm caused by incorrect DCMA request.

Just because a video is fair use doesn't mean that youtube has to host it. It is a reasonable business decision for them to make to just block all of it.

I am somewhat confused as to why brief clips interspersed with other material wouldn't be automatically marked as fair use and left on their own.

The answer to all this is to simply post nothing but obnoxious gruel at all times. Spam endless amounts of sarcastic nothingness shot through with zombified imitation teenage drivel.

Nonsense so unbearable that it enrages humans on the other end, so they consider suicide, as they wade through this drek which they must now consider as possible violation in a civil dispute.

There are humans in this loop. Bet on it. And make them pay.

What do people think of this video? https://www.youtube.com/watch?v=t7tA3NNKF0Q

IMO Content ID is essential for a healthy video creator ecosystem. Google's implementation may give the big media companies too much power, but for indie creators it is crucial. Facebook, on the other hand, seems to not care about video piracy at all and as a result small creators suffer.