This is another symptom of Google's extreme "let's make algorithms do what people usually did". In an effort to automate everything and minimize employee-customer interaction, Google creates automated systems that work correctly 90% of the time. The 10% of the time it doesn't work correctly, you're fucked, and there's no way to contact anyone at Google to dispute or fix things.
Google just doesn't care at this point, because trying to fix the remaining 10% apparently costs more than the resulting good user experience / satisfaction would generate. This is also very visible in AdSense and Gmail. I think Google just doesn't like edge-cases.
This is another symptom of Google's extreme "let's make algorithms do what people usually did".
Actually humans can't process all the video uploaded to Youtube, and never did. Viacom sued them for $1,000,000,000 (billion) and more rights-holders were lining up behind them to do the same. Not just for money - a lot of very powerful companies wanted Youtube shut down for good, and were willing to throw real resources at making that happen. Google put ContentID in place to placate the rights-holders in order to save Youtube's existence. It would literally be sued into oblivion and no longer exist if it weren't for ContentID. So there are issues, but don't think Google just did this on a whim.
Of course not, but once you have edge cases like a dispute, there needs to be a human who can view both videos and check who is right, instead of just starting another automated system. There needs to be a system in place to punish those who wrongfully claim videos, and that probably needs human overview.
The DMCA, flawed as it is, would have saved YouTube's existence. They created ContentID to mollify the Content industry into playing nice and creating the dumping ground that is Vevo and letting their products appear in the Google Play store. Without Content ID, they surely would not be playing along with Google's media store plans.
YouTube the site is clearly and plainly within DMCA safe harbor and has absolutely no need for ContentID to continue operating. It is the other things Google wanted to do with big Content that required this asinine system.
I'm reaching the point where that argument no longer holds water for me. So it's impossible to run a video site as large as Youtube without breaking the law or screwing people over? Then don't do it! Would the world really be worse off if instead of Youtube, we had 1,000s of much smaller sites, capable of handling their workload with real actual people?
Is the world really worse off with Youtube the way it is? or even the way it was, without ContentID? I can't say I see the harm in some people uploading crappy copies of some videos and not making a profit on them. Content owners will always overreact even if nothing bad is happening, for example this just happened: http://yro.slashdot.org/story/12/08/09/007212/legitimate-ebo...
> Is the world really worse off with Youtube the way it is?
With videos of Curiosity being taken down for copyright infringement? With "the little guy" having to wait up to a month for anyone to prove wrongdoing, and in the meantime their video goes offline? I'd say that's pretty bad, yeah.
I understand why they do this to regular people, but I don't understand why they don't whitelist accounts like NASA's so that they don't run into these problems.
Agree. I have an extensive background with Mechanical Turk, and this is a perfect problem for such the service. Google, I'm volunteering my time here to refine the DMCA takedown process.
I am a Content ID publisher - the point of it is, if you sign up with YouTube as a partner, you commit to not upload content that you didn't produce (in so many words).
So in an ideal world, every piece of content from a Partner would be unique and original, and Content ID would work as intended.
The key is - 100% unique and original - it's not original if it's 2 anchors introducing a clip of the Mars landing.
The problem here isn't just YouTube - it's partners that are uploading content that they did not entirely create.
The system only works if people cooperate; if news orgs contribute original content and claim it (and other uploads) everything is ok - the problem is that news orgs are just dumping content on to YouTube without thinking about that.
We accidentally published a movie review to YouTube and CID claimed it, and for the next 3 months I was constantly releasing claims, until I finally unclaimed our own movie review completely to stop the matching.
TL;DR - publishers are very much at fault, as much or more as YouTube is.
I don't see how. From the publisher's point of view, it's completely legit: A video that mixes public domain imagery with original additions. Any news org that covered the landing without including some of NASA's public domain images or video would have everyone (rightly!) saying "Their Mars coverage sucks!"
Even if they don't have original additions, it's in the public domain; publishers can do what they want with it, including merely slapping their watermark in the corner and uploading as-is.
I'd say that nobody's really "at fault;" rather, the widespread availability of technology to easily and cheaply download, remix, upload, and stream video is presenting new use cases which our existing notions of content ownership simply haven't had to deal with before.
The solution will be through some mix of better technology (i.e. smarter scanning algorithms), better modeling of the problem domain (maybe the next version of YouTube's Content ID program will address the issue of situations like this), and changing social and legal norms.
In the long term, I believe that one day, people, businesses, laws, and private agreements like ToS will work out an equilibrium where expectations of ownership and use are clear to everyone, penalties are considered reasonable and proportionate to offenses, are enforced consistently with few false positives or negatives, and people mostly manage to get along -- in contrast to the current free-for-all environment where none of these things are true.
Unfortunately, "the long term" may be very long -- I'd say 30-60 years. Also, while my assessment may sound positive on first reading, the previous paragraph actually leaves substantial space for very distopian scenarios -- where (for example) permanent ownership is considered reasonable, fair use no longer exists, the expectation is that copyright violators will be jailed, financially ruined, and/or permabanned from the Internet, those penalties are swiftly and surely enforced by extremely invasive monitoring of all computer activity, and anyone who tries to make waves about the situation is dismissed by the media establishment as a fringe lunatic, and can't effectively organize due to the intense restrictions on the Internet.
> I don't see how. From the publisher's point of view, it's completely legit: A video that mixes public domain imagery with original additions. Any news org that covered the landing without including some of NASA's public domain images or video would have everyone (rightly!) saying "Their Mars coverage sucks!"
Why does the stuff they upload to YouTube's content violation detector include public domain content? It doesn't have to be, and shouldn't be, exactly the same as what they broadcast.
Exactly - publishers shouldn't upload and claim content with public domain footage in it; they should specifically remove any videos with content not originally produced by the TV station from the CID system. But that takes work, and generally people just shovel content from TV -> Web, let CID make the decisions, and move on. It's one mouse click to turn off CID matching, really easy.
It's YouTube's responsibility to implement a system that doesn't defraud people. The users getting ripped off have no direct relationship with the other users who are abusing the system, and no way to hold them accountable.
Are CID publishers subject to claims from other CID publishers when they both wrongly claim the same video (like the rover landing?) Or are there first and second class citizens on Youtube, with the first class citizens not subject to such things?
I've never seen a claim from another partner, I think it's first-in gets the claim, then others can dispute the claim. Like I've had a video uploaded by someone else that we own start to go viral, so we upload a fingerprint to get the match, then make a claim against that user-upload, which someone else may have already tried to claim. It's really complicated.
The problem is that its algorithmic. The content owners aren't clicking around and claiming videos -- Google's Content ID system finds videos that look similar to theirs and then claims the video.
The ideal case would be someone reposting a trailer or a music video to Youtube. The content is identified as being owned by a partner. Now the partner gets the money and the poster still gets views.
Cases like the OP thus cause a problem: the majority of the content is 3rd party and free.
Note: I'm not defending CID or saying its a good thing. I just think your proposed solution won't fix the issue at hand.
The initial claiming process is automatic. But as noted in the article, the uploader then has an opportunity to dispute the claim. The rights holder is then supposed to manually verify it and either approve or dismiss it. The problem is that there is no penalty for just approving every claim after "reviewing" it. Contrast this to the DMCA process, where after the uploader files a counter claim the rights holder must file a court order in order to keep the material down.
At the moment it's a terribly one-sided process that allows big media companies to profit for absolutely no reason. Fining companies for approving disputed claims in error seems like a pretty good solution.
There's no abuse mentioned in this particular article, and in this article it is exactly a "technologically broken" issue.
The article says that if a TV network's real, legitimately copyrighted feed, that they are correct in telling Youtube is copyrighted uses public domain video in some parts (which doesn't mean their particular feed based on it isn't still copyrighted), then other videos that use the same public domain video (from the original source, not off of the copyrighted feed) are identified as being the copyrighted source.
YouTube has become a Kafkaesque nightmare, at least for regular users. They recently opened up monetization to everyone, but if you actually try to use it, half your videos will go "under review" and you will be asked to provide proof that you own all the content. They won't tell you specifically what you have to provide, and nobody I know has managed to figure it out or get a video out of this state. Of course, there is no way to contact anyone or get any more information.
Apparently, "full partners" can actually get real customer service and avoid these problems, but you need to get on the order of 1000 views/day before they offer you that, which most people never achieve.
I will absolutely never do business with that company.
It seems like when you submit material to the content ID system, the first thing it should do is check it against material that's already been registered. Conflicts found there need to be resolved before that material is used to automatically send out ownership claims.
This change would surely weed out a big chunk of these false claims, and would put a lot of the effort for doing so squarely on the shoulders of the claimants.
Hmmm could this be used to turn the tables. What I'm thinking is you buy a area of land that needs clearing up - video it and when it is claimed as copyright you then sue the copyright abuser to clear up the land as they claimed its there rubbish. Not best example but hopefuly shows how this can be levelled.
What has happened is alot have loaded up alot of common things and sounds and as such any autobot will match them up with pretty much anything you do in some situations.
Can bet if you upload a video of your first born that it will match up with something somewere else. It's like patent trolls, only automated, utter nightmare.
This is also one of those moments were I liek the UK libel laws as deformation of character under the guise of falsly breaking copyright of somebody else's contents is one of those area's were you would have fun in court and fiscaly rewarding.
Google just doesn't care at this point, because trying to fix the remaining 10% apparently costs more than the resulting good user experience / satisfaction would generate. This is also very visible in AdSense and Gmail. I think Google just doesn't like edge-cases.
Highly relevant Dilbert comic: http://dilbert.com/strips/comic/2001-02-25/