The MTTLR Blog

Editor: Part 1 introduced the emergence of filtering as a potential solution to copyright and infringement, and examined why it fails copyright owners. This part considers the problems of filtering from the perpsective of users and Google.

2) Users

YVI is also a terrible blow to utilize who use YouTube to upload their creative efforts as well as to the Internet community at large. Although YouTube has pledged that YVI will not “impede the free and fast communication YouTube has enabled,”²³ it has yet to explain how it will allow for fair use of video material.²⁴ Here’s how the system works: When a user uploads a video, that video is run through YVI to see if there’s a match with the content in its database. If so, the video is subject to the action the rights holder has decided to apply to it—it could be blocked, “tracked” or “monetized.”²⁵ If it is blocked, the user can contest the decision and YouTube can put the video up. It is at that point that the DMCA takedown provisions kick in. If the copyright holder wants the video taken down, it can send a takedown letter to the user (as it would under the old system).²⁶

This is problematic in two ways. Despite YouTube’s claims regarding “free and fast communication,” it will almost surely sweep in an unacceptable level of false positives.²⁷ As a result, it will have the effect of stifling expression. YouTube is a haven for users who use copyrighted content to create new forms of expression,²⁸ and the technology would have to be sophisticated indeed to not block these otherwise protected forms of expression. Of course, YouTube is not required to post a video just because it doesn’t infringe a copyright. Indeed, even if YVI sweeps in what would otherwise be protected video, YouTube may delete videos it even suspects of infringement. However, to do so would be to ruin what has made YouTube great for the online community: the democratization²⁹ of creative expression by creating a repository for content that anyone can use. If faced with a notice from YouTube that their uploaded video may be infringing,³⁰it is likely that many users may not be willing to “put themselves in the crosshairs of movie studio lawyers”³¹ and may instead decide to keep their expression to themselves.³²

Electronic Frontier Foundation’s Fred von Lohmann has proposed two modifications to the system that would help protect fair users. The first is to require a video and an audio match before automatically blocking content.³³ The second is to determine what ratio of the uploaded video is comprised of protected content—if the ratio is low, it is likely the post counts as transformative content.³⁴ This is a step forward, but still leaves a lot of protected expression up in the air.³⁵ Another possible solution—to add a human component to the review process—simply won’t do. YouTube is likely unwilling to open itself up to the liability that would occur if one of its reviewers green-lights a video that turns out to be infringing. It would also be too costly and time-consuming to personally review every movie that gets posted on YouTube. The best solution, it seems, is to continue to place the burden on copyright owners to find what they believe is copyright infringement and to allow them to follow the DMCA’s procedures. YVI places an extra burden on fair use that neither the courts nor Congress has required.

3) Google

YVI is also an unnecessary and unfortunate step for Google because it does more than the law requires.³⁶ A compelling argument can be made³⁷ that YouTube falls well within the DMCA safe-harbor provisions.³⁸ That is, even without YVI, YouTube could very well escape liability because of its substantial compliance with the DMCA. Viacom’s complaint is, at its heart, that YouTube is following the law too closely for its liking.³⁹ Google’s decision to implement YVI is too large a concession to make to the copyright owners—Google is now the one responsible for finding infringing videos (and for creating and maintaining the software responsible), even though the DMCA places that burden on the copyright owners. Paradoxically, YVI could actually be used against Google: implementation of YVI shows that Google is well aware of the copyright infringement that exists on its website. By providing the filtration service, Google may be shouldering more responsibility than it’s prepared to take on. If YVI fails to flag a protected video, Google will have a harder time taking refuge in the DMCA’s safe-harbor provisions because the knowledge requirement⁴⁰ clearly will have been met.

It is likely that, feeling the pressure from Viacom’s lawsuit, Google wanted to show that it was serious about copyright infringement on its website. YVI was a poor way to do this for three reasons. First, although the Supreme Court has implicitly favored attempts at filtration,⁴¹ it has by no means required it.⁴² Without other evidence of intent that the product be used to infringe, Grokster implies that filtration is unnecessary (although concededly helpful). Second, by implementing the software, Google may “have the practical effect of changing filtering from ‘one’ factor to ‘the’ factor that a court considers” in deciding this sort of infringement action.⁴³ Because of Google/YouTube’s size and market presence, its decision to create YVI could likely become the industry standard. If that occurs, it will make it more difficult for smaller websites to host content because of the costs involved in developing or licensing filtration software. Third, YVI was a poor response to the Viacom lawsuit because it simply will not help Google fight claims of past infringement.⁴⁴ Instead, this move looks like a signal of YouTube’s imminent decline. As copyright owners begin to develop their own content-distribution systems,⁴⁵ it seems inevitable that users will begin to migrate to the websites hosting the popular content.⁴⁶ YouTube became the market leader due, in part, to its hosting of copyrighted content.⁴⁷ Without that content, YouTube must tread very carefully to avoid losing the users that have made it worth $1.65 billion. Implementation of YVI is a step in the opposite direction.

4) Conclusion

YVI is a bad solution for all parties concerned. I do not mean, however, to overstate the case against YVI. Undoubtedly, it will have a marked effect on cutting down on clearly infringing videos posted on YouTube, thus decreasing the number of separate acts of infringement against copyright owners. Instead, my argument is simply that YVI will not be completely effective against copyright infringement and that it is no greater a solution than the licensing system already in place.

It is also not the case that YVI will inevitably lead to the total downfall of YouTube: YouTube is popular, in part, because it is a common home for a wide range of wholly original works. YVI will surely not have an effect on that aspect of use. However, many YouTube videos incorporate copyrighted content into original works of creative expression. If one accepts the proposition that a great number of those works should properly be considered fair use,⁴⁸ then YVI will end up imposing a large burden on this legitimate form of expression. Restrictions on the ability of users to use copyrighted content in the works they upload to YouTube will harm YouTube’s business model and the Internet community at large. Indeed, I believe that any undue difficulty in getting a legitimate fair use video past the system will have the ultimate effect of user migration.

This is not to say that it’s a bad thing to keep infringing works off the web: obviously, if a work is infringing it should be flagged and removed. I only mean to argue that YVI flags too much while, at the same time, failing to safeguard the interests of the copyright owners it was implemented to protect.

The growing emergence of “Web 2.0”¹ has caused an explosion in the amount of user-generated content that appears on the Internet.² YouTube.com has revolutionized the way we watch some of this content online, by allowing users to upload video files stored on their computer and by combining social networking features with that video. Amazingly, over six hours of video are uploaded onto YouTube every minute.³ The success YouTube has enjoyed over the past few years has not been unnoticed—its user base is so lucrative that last year Google purchased YouTube for the princely sum of $1.65 billion.⁴ Given the activities of those users, however, YouTube has found itself the target of numerous lawsuits⁵ by copyright owners (the deep pockets of its new investor provide an additional incentive). Perhaps most notable is the lawsuit brought by Viacom: tired of filing DMCA takedowns for the over 150,000 unauthorized clips it claims are hosted by YouTube,⁶ Viacom is asking for $1 billion in damages for copyright infringement,⁷ despite YouTube’s removal of over 100,000 Viacom-owned videos in February.⁸

Viacom’s consternation is based, in part, on YouTube’s withholding, until now, of filtering technology that would ease Viacom’s burden of issuing DMCA takedown notices for infringing YouTube videos.⁹ YouTube has dragged its feet in implementing this technology claiming technical problems¹⁰ (it’s worth noting that YouTube has had no problem in developing a process that keeps pornographic videos from being uploaded) but did roll out a beta version of its new “YouTube Video Identification” (hereinafter, YVI) software just last month.¹¹ YVI is the newest of several policies and tools YouTube has in place to limit the level of copyright infringement that takes place on its website.¹² This new software, designed to facilitate the “free and fast communication YouTube has enabled”¹³ has been criticized by copyright owners as “too little, too late”¹⁴ and by copyright watchdogs as substantially overbroad.¹⁵ I would argue that not only is YVI a poor solution for both groups, it is also an unwise strategic move by Google.

1) Copyright Owners

Copyright owners like Viacom should be careful what they wish for—YVI shows they just might get it. Unfortunately for them, YVI is hardly the solution that copyright owners were hoping for. Although it may seem like a “logical way to try to balance the needs of users and copyright holders,”¹⁶ it actually imposes excessive burdens on copyright owners that will likely render the software ineffective. Cory Doctorow argues that it would be impossible to write a program that is able to spot all copyrighted works, as well as every “transformation, re-encoding, downsampling, and re-edit of those works.”¹⁷ If it is unable to do so, a large number of copyrighted works will fall through the cracks. This is practically guaranteed, when one considers the sheer amount of time and money it would take for a corporation like Viacom to turn in a copy of every copyrighted work it owns to Google.

Technological limitations aside, the program simply asks too much of copyright owners. YouTube wants the “content community” to “help [YouTube] help [the copyright owners]”¹⁸ by handing over to Google all of the copyrighted material they wish to protect. Google will then use the content repository to find works on its website that match up with the protected works in its database. By placing the onus to produce on the copyright owners, Google can claim that it’s not their fault if their filtering software failed to catch the infringing content if the protected work isn’t in Google’s massive database. At the same time, copyright owners are extremely reluctant to hand over such a large amount of material to “the ultimate big brother.”¹⁹ Indeed, the system essentially establishes a two-tiered copyright system with respect to YouTube: those that have the means and the willpower to, in effect, register their copyrights with Google²⁰ will receive protection, while those that fail to register will receive nothing. This does nothing to alleviate copyright owners’ concern that the current system, whereby YouTube offers filtration protection only to companies that sign licensing agreements, is also a two-tiered copyright system divided between those that play ball with YouTube and those that don’t.

One solution to this problem, proposed by Viacom, is to set up an industry-wide content repository instead of allowing content providers like YouTube to continue to develop proprietary systems of their own.²¹ One way to do this might be to add to the 1976 Copyright Act’s registration requirement: require also that a copyright holder place on file with the Copyright Office a digital copy of the protected content in order to commence an infringement action. The Copyright Office could then allow third parties like YouTube to use their filtration systems to match up posted content on the website with registered content at the Copyright Office. This would eliminate the “whack a mole” problem that currently plagues corporations like Viacom—as more websites roll out their own unique proprietary content filtration systems, copyright owners like Viacom are forced to keep up with each system’s requirements by handing over their content to each one.²² Of course, copyright owners might be nearly as averse to creating a database of content owned and controlled by the government as they are to creating one that Google controls. Still, with proper congressional oversight, the program seems more likely to win support from the corporate community.

Editor: Part 2 will consider the problems of filtering from the perpsective of users and Google. It will publish later today.