YouTube’s Content ID

There’s something about the nuances of copyright law that fascinates me, so I was very excited to be able to take IS 289: Intellectual Property Law for Librarians and Archivists in the fall 2014 quarter. While I was originally planning on writing about the Digital Millennium Copyright Act (DMCA), my research began turning up many instances of individuals having their own materials removed as a result of supposed copyright infringement due to YouTube’s automatic copyright system, Content ID.

Access to digital moving image technology has revolutionized the ways in which people create media content today, and this surplus of materials–and access to copyrighted materials–has greatly affected copyright law. Content ID essentially threatens an individual’s ability to claim fair use and, copyright law, much like archival science, is not widely understood by typical media users. In this paper, I sought to simplify and analyze copyright law as it pertains to Content ID in order to education media creators on the effects of systems such as Content ID on fair use. 


YouTube’s Content ID: Balancing Copyright Protection and Fair Use on the Web

YouTube Content Copyright Violation
Photo from Forbes.com
Abstract

Videos are added to YouTube at the rate of 100 hours per minute.[1] This vast amount of content uploaded to the web by users raises concerns about the ability to control the illegal spreading of copyrighted works. With the implementation of Content ID, a system that checks for copyright infringement by comparing all uploaded materials against a database of copyrighted works, YouTube is seeking to create an automated way to prevent copyright infringement on its website. This automation has its drawbacks and can, at times, be misused or abused in order to violate one’s ability to claim fair use. Content ID and similar systems provide copyright protection but circumvent the DMCA, and could be detrimental to the safe harbor provisions by making it standard practice to implement such systems, rather than rely on the DMCA safe harbor for protection.

Introduction

In an increasingly digital world, where content is only getting easier to share and copy, there is a need for programs such as YouTube’s Content ID system. However, the implementation of these programs should take care to not negate the DMCA takedown process or infringe upon the right to fair use. It is not difficult to discover cases in which Content ID has unjustly blocked users’ videos, some videos that contain only the original creative work of the uploader. As there are no clear lines determining what constitutes fair use and what doesn’t, automated systems such as Content ID do not have the ability to determine that the use of a copyrighted work is fair use or not. The system can only accomplish what it has been created to do, and that is to detect copyrighted works in videos that have been uploaded. Thus, Content ID cannot be relied upon to be the ultimate authority of whether a work should be allowed on YouTube or not. YouTube has failed to provide users with a meaningful way to appeal Content ID matches, as it relies on the rights holders to help regulate the fair use claims.

What Exactly is Content ID?

Released by YouTube in 2008, Content ID is a system that scans all videos uploaded to their website and checks the contents for infringing materials. In order for this to work, YouTube allows copyright holders to submit copies of material that they own the exclusive rights to. These reference copies are stored in a database and used to compare the uploaded materials with the copyrighted materials.[2] YouTube will not post videos that turn up as a match, giving the copyright holder the option to block, mute, or monetize any videos containing their content. Users are allowed to submit an appeal, should their videos be improperly flagged or if they believe they qualify for fair use.

As YouTube users upload 100 hours worth of video material per minute[3], programs such as Content ID offer a way to control the amount of copyrighted materials that gets uploaded illegally, preventing rights holders from having to manually sift through videos to ensure that their material in not being infringed. Content ID is not available to just any user that owns copyrighted material, however. YouTube allows rights holders to use the system “based on an evaluation of each applicant’s actual need for the tools.”[4] Actual need is not defined.

For users who do not qualify for Content ID, or as YouTube states, for users whose needs would be best served through other copyright protection tools,[5] there are alternate ways to request the removal of infringing content. These tools include a copyright notification web form, which is essentially a DMCA takedown that allows a user to submit a form for a specific video that infringes their copyright, and the YouTube Content Verification Program (CVP), which allows users to select multiple videos at a time, and request takedown notices for those selected videos. Unlike either of these options, Content ID is an automated system that does not require rights holders to target specific videos.

Issues with Content ID
YouTube Content ID
Photo from lazygamer.net

While Content ID has helped to eliminate the amount of copyrighted material that is being illegally uploaded to YouTube, it has a negative effect on those who are using these materials under fair use and whose videos contain materials they have legally licensed. Errors that arise can involve the misidentifying of content, forced advertising for videos with matched content, and a difficult appeals process. All of these problems have caused a great uproar in YouTube’s user communities, and have specifically targeted certain communities. Gamers who use YouTube to share videos of themselves playing certain video games often times have the blessing of the rights holders of the games. In December of last year, Content ID identified many of these videos and they were removed from the site, mostly due to copyright issues concerning the music that the games contain. In response to the outrage this caused, YouTube issued an apology that did not sit well with many of the angry users.[6] The letter informs users of the issues with content that could contain multiple rights, and advises them to mute any music that could possibly belong to a third-party rights holder to ensure their videos do not get flagged. Despite the outrage of the gaming community over this letter, the music in the background of those videos may very well be infringing and advising users to be wary of this content is not an unreasonable request. Nonetheless, this instance highlights YouTube’s lack of transparency with regards to Content ID and its inability to take into consideration the specific needs of a large group of users, who are quick to point out the many injustices that they have faced due to Content ID. Many of the problems that have emerged with Content ID could have been prevented, or been reduced, had YouTube taken into consideration the needs of its users.

1. Program Malfunctions

One issue with Content ID revolves around the abilities of the program itself. Tests performed by YouTube users determined that earlier versions of Content ID were faulty and could be easily tricked. In 2009 blogger Scot Smitelli ran a series of tests, uploading a copyrighted song with various forms of audio modifications in order to determine the limits of the system. One of the more interesting discoveries Smitelli made was that when the first 30 seconds of a song was muted, Content ID did not catch it.[7] A variety of other sound modification tests proved that the software behind Content ID was certainly not foolproof and was, at times, rather arbitrary in what it would allow and what it would block. Since the time of Smitelli’s blog post, Content ID has been updated and the results of the test would certainly be very different today.

Content ID has also been known to mistakenly identify users’ videos as containing copyrighted material. Perhaps the most egregious instance of this was when YouTube user TanRu Nomand had a hand-coded video game, that he created himself, taken down for supposedly containing third-party content.[8] This particular instance happened just a year ago, in December 2013, proving that despite updates to the system, Content ID can still be unreliable. There are a multitude of examples of copyright fraud and rights holders claiming material from videos as their own when, in fact, they are actually the original, creative property of the uploader.[9] With Content ID, rights holders are meant to review each appeal that is made when their content is identified, however, the system makes it easy for rights holder to ignore this, and simply claim everything as their own.

2. Monetization and Claiming
Others’ Intellectual Property

            Content ID is designed to inform the rights holders when something has been flagged for containing their content. The rights holder then has the ability to block the video, or to allow it to stay on YouTube while monetizing it by running ads. This provides a compromise that allows users to upload copyrighted materials, which the rights holder can profit from, through a sort of internal licensing agreement by way of YouTube. While this may appear to be a good system for dealing with copyright infringers, it has proved to be problematic for those users who are using legally licensed content in their videos.

The owner of the WatchMojo YouTube channel details how the Content ID copyright detection program took down the WatchMojo channel, and required him to appeal the site’s videos in order to get them back online. The case of WatchMojo brings up the issue of Content ID affecting YouTube users who use the site as a way to earn income. In some cases creating YouTube videos can be a full time job and Content ID claims can have a negative effect on their income. When content is flagged, and content owners are allowed the option of monetizing content, this can detract from the income of those who rely on YouTube to make a living and who are complying with copyright laws by way of fair use or through licensing. As the WatchMojo channel owner points out, “Why does YouTube treat (in regards to ContentID) a 13-year old’s channel almost the same way it treats the channel of a media organization?”[10] Regaining the rights to monetize your own content after it has been flagged is a very daunting and difficult task, and even if you are allowed to monetize your own videos the rights holders could chose to share their revenue with you, diverting some of your income. As of now, there is no way to show that the content you are attempting to upload has been legally licensed from the rights holder. Thus, in order to upload any content in which you have the right to use, you must first go through the YouTube appeals process.

3. Issues with the Appeals Process

To correct any issues that may arise with the Content ID system, YouTube implemented an appeals process for users who believe their videos were unrightfully removed. This process employs a guilty until proven innocent system, where videos found to contain third party materials are blocked until they have been successfully appealed. Thus, users uploading videos with legally used content or videos that qualify for fair use, are at the mercy of the rights holder before their videos can be viewed.

Perhaps one of the key issues with regards to the Content ID system has to do with what happens when a user appeals a content match: that appeal goes to the rights holder for review. In the early days of Content ID, a rights holder could then deny the claim, and that was the end of the matter. It wasn’t until YouTube revised their appeals process in 2012 that rights holders had to either release a claim, or formally file a DMCA takedown notice.[11] The system formerly in place allowed rights holders to censor and control the way their content was being viewed online. If they didn’t like the way someone was using their material, they could simply have it removed and the user who uploaded the video would have no way to challenge that decision. There are legal consequences for improper use of DMCA takedowns, due to the Lenz v. Universal ruling, which requires the rights holder to take into account fair use before issuing DMCA takedowns[12] and encourages rights holders to ensure that they are not abusing their privileges. However, Content ID blocks have no such consequences.

Content ID Message
Photo from youtube.com

Besides these issues with the appeals process, YouTube has also instated a limit on how many claims can be disputed at a time. YouTube users can only counter Content ID claims three videos at a time.[13] While this can be beneficial to prevent unashamedly infringing users from disputing multiple claims, it presents an issue for other users, such as WatchMojo, who was subject to a weeks worth of repairs to fix the damage caused by Content ID.

Viacom v. YouTube

The DMCA gives rights holders a process by which to have their infringing materials removed. The system in place, while it may have its issues, has been known to work. Why, then, would YouTube go to the trouble to make Content ID? Clearly the pressure from rights holders, undoubtedly very large media companies with multitudes of assets, has convinced YouTube that such a system is necessary. In the years before Content ID, YouTube had a shaky relationship with these media companies.[14] The lawsuit Viacom v. YouTube brings up many of the issues at play and demonstrates the power struggle between two immensely large companies.

In 2007 Viacom brought a one billion dollar lawsuit against YouTube, and parent company Google, for copyright infringement, claiming that YouTube was aware of and encouraged the uploading of copyrighted materials to its website, and was profiting from these materials.[15]

Under the Digital Millennium Copyright Act of 1998 third-party hosting websites are protected from lawsuits brought against them based on what their users post to their site. To qualify for this safe harbor protection, however, the websites must meet a variety of criteria, which includes being unaware of the existence of copyrighted materials on their site. The charges against YouTube claimed that they were in violation of aspects of 17 U.S.C. § 512(c), specifically that YouTube had “actual knowledge” of infringing materials and were “aware of facts and circumstances from which infringing activities is apparent.”[16] However, working in YouTube’s favor, it was acknowledged that because of YouTube’s immediate removal of material that Viacom had issued a takedown notice for, they were complying with 17 U.S.C. § 512(c)(1)(A)(iii), “upon obtaining such knowledge or awareness, acts expeditiously to remove, or disable access to, the material” and 17 U.S.C. § 512 (c)(1)(C), “upon notification of claimed infringement … responds expeditiously to remove, or disable access to, the material that is claimed to be infringing or to be the subject of infringing activity.”[17]

Citing these DMCA’s safe harbor provisions, District Judge Louis L. Stanton granted YouTube’s request for summary judgment.[18] Viacom appealed, on the grounds that a jury could find that YouTube was guilty of copyright infringement, and brought up a number of emails that had surfaced during the lawsuit, in which YouTube executives acknowledged the presence of infringing material. Google, in turn, brought up some of Viacom’s own email correspondence, showing that they had secretly uploaded their own content to YouTube, which they soon lost track of.[19] Some of this content that was removed by Content ID, and Viacom had to ask YouTube to reinstate it. Based on this new evidence the judge made the decision that YouTube, while aware of the existence of copyrighted materials on its site, was not expected to know which of these works were infringing and which were licensed and, therefore, used legally. YouTube was once again granted summary judgment, and Viacom was in the process of appealing for a second time when the two companies settled out of court in March 2014.[20] It is unknown what the terms of the settlement were.

The Viacom v. YouTube lawsuit brings up some important issues about the DMCA that the implementation of YouTube’s Content ID only makes more complicated. The initial judgment on the case proved that the DMCA takedown notice system that was in place was successful, and proved the effectiveness of the safe harbor provisions. In a New York Times article written after Judge Stanton initially granted summary judgment to YouTube, former Google lawyer Michael S. Kwun said in an interview that the decision would prevent companies from being legally required to develop programs such as Content ID, which would be cost-prohibitive enough that many start up companies would not stand a chance. Despite the apparent win for the DMCA, Viacom was no longer seeking damages for copyrighted material uploaded after the implementation of YouTube’s Content ID system in 2008.[21] Ultimately, Viacom’s persistence in continuing to appeal the case seems to be a warning to smaller companies that attempt to build a YouTube-like content sharing system. YouTube, backed by its parent company Google, was able to fight Viacom in court, but any smaller company or start-up may not have the financial resources to do so. While the initial summary judgment ruling may have been a considered a win for these smaller companies, Viacom’s persistence in pursuing the lawsuit, proved that creating a website similar to YouTube would be a difficult task for any smaller companies in the future.

 Similar Programs

Another popular Internet video steaming service, Vimeo, launched its Copyright Match Program in May 2014. Vimeo’s Copyright Match uses the program Audible Magic to scan through all content uploaded on their website.[22] Unlike YouTube, they took a very different approach in how they have implemented their system, beginning with their announcement to the user community. Vimeo has been very up front about how their system will work and has listened to user feedback to determine how to make the changeover. Despite Vimeo’s open and transparent implementation of Copyright Match, they have still received a great deal of backlash from their users. The Vimeo Staff blog post, where the Match program was announced, is filled with comments from angry and disappointed users, who fear that Vimeo will fall prey to many of the same mistakes that YouTube did when it first rolled out Content ID.[23]

One of the major differences between Vimeo and YouTube with regards to their automated copyright detection systems is that Vimeo has a more refined appeals process. As already stated, YouTube has a tendency to assume infringement before proven otherwise, and requires the rights holder to approve of the use of its own material, all the while keeping the video blocked. On the other hand, Vimeo states that they will leave the video up, unless the appeal is denied. Their appeals processes appears to be far more user friendly, encouraging users to argue their fair use case, or to prove their right to use the content via licenses. Another very importance difference is that Vimeo staff, not rights holders, review the appeals. This in itself decreases bias in the appeal process and allows for a more objective implementation of fair use.

Content ID and the DMCA

Lenz v. Universal Music Group upheld that rights holders must take into account fair use before issuing DMCA takedown notices.[24] Content ID, not being directly associated with the DMCA, negates this lawsuit through the automated removal of materials that may qualify for fair use. While users do have the option to appeal the decision, this may deter those who are unfamiliar with the Content ID appeals process, or those who are unfamiliar with intellectual property law and may be afraid of negative repercussions should they appeal. While Content ID proves to be a very valuable resource for protecting the rights of copyright holders, this resource is not without its flaws. Often this protection is at the expense of users’ rights of Fair Use and Freedom of Speech.

The prevalence of Content ID, and the fact that other video sharing services are adopting similar programs, is troublesome. As stated in a New York Times article, the initial ruling of the case proves that start-up companies will not have to build up systems such as Content ID.[25] However Viacom’s appeal contradicts this entire comment. Websites may be forced to turn to this technology, as new companies will not have the resources that Google does in order to stop lawsuits such as the one Viacom filed against YouTube. If the practice of using programs such as Content ID becomes widespread, or at least necessary to protect against expensive lawsuits, will it hinder the growth of smaller or start-up companies?

YouTube Error
Photo from mashable.com

Content ID also brings up the issue of how involved web service providers must be in circumventing copyright infringement. DMCA grants a safe harbor for companies that are not aware of the existence of infringing content, however, any website that allows users to upload material is bound to contain such content. While Viacom v. YouTube proved that there would be protection for companies in this instance, it may only be a matter of time before rights holders demand that such websites take more control over what is posted to their sites. Content ID sets a standard that other companies are obliged to follow, should they wish to protect themselves against lawsuits. This leads to a question of whether or not the DMCA will end up being completely annulled by companies taking copyright issues into their own hands in the way that YouTube has. Will there even be a need for the DMCA in the future?

Conclusion

As it stands now, Content ID is being used as a substitute for the DMCA takedown notices and, thus, functions outside the realm of government regulation, giving YouTube the chance to regulate itself. Hopefully, this will not become a necessary standard practice in third-party hosting websites such as YouTube and Vimeo, and the DMCA will continue to remain effective in safeguarding these companies. However, as more and more content is added to the web each day, tools such as Content ID will only grow more prevalent and will need to be constantly refined. It will be interesting to see how YouTube continues to deal with the issues that result from abuses with its system, and if it will learn to be more transparent with its changes. While there may be a need for systems such as Content ID, companies will always have to balance protecting their rights holders with the needs of their users.

Works Cited

[1] “Statistics – YouTube.” Accessed December 12, 2014. http://www.youtube.com/yt/press/statistics.html.

[2] “How Content ID Works – YouTube Help.” Accessed December 12, 2014. https://support.google.com/youtube/answer/2797370?hl=en.

[3] “Statistics – YouTube.”

[4] “Qualifying for Content ID – YouTube Help.” Accessed December 12, 2014. https://support.google.com/youtube/answer/1311402.

[5] “Qualifying for Content ID”

[6] “Here’s YouTube’s Reply To Angry YouTubers About This Content ID Mess.” Kotaku. Accessed December 12, 2014. http://kotaku.com/heres-what-youtube-has-to-say-to-angry-youtubers-1485168478.

[7] “Fun with YouTube’s Audio Content ID System • Scott Smitelli.” Accessed December 8, 2014. http://www.scottsmitelli.com/articles/youtube-audio-content-id.

[8] My Own Game Has Been Flagged by YouTube!, 2013. https://www.youtube.com/watch?v=wJKhGl5DIDU&feature=youtube_gdata_player.

[9] “YouTube Copyfraud & Abuse of the Content ID System.” Accessed December 15, 2014. http://fairusetube.org/youtube-copyfraud.

[10] “It’s Time For YouTube’s ContentID & Copyright Policies To Reflect Reality.” It’s Time For YouTube’s ContentID & Copyright Policies To Reflect Reality. Accessed December 12, 2014. http://www.tubefilter.com/2013/12/23/youtube-contentid-reflects-reality/.

[11] “Official Blog: Improving Content ID.” Accessed December 13, 2014. http://youtube-global.blogspot.com/2012/10/improving-content-id.html.

[12] “Lenz v. Universal.” Electronic Frontier Foundation. Accessed December 13, 2014. https://www.eff.org/cases/lenz-v-universal.

[13] “It’s Time For YouTube’s ContentID & Copyright Policies To Reflect Reality.”

[14] Hansell, Saul. “YouTube’s Video Poker; A Bet That Media Companies Will Want to Share Ad Revenue.” The New York Times, September 30, 2006. http://query.nytimes.com/gst/fullpage.html.

[15] “Viacom International, Inc. et Al v. Youtube, Inc. et Al, 1:07-Cv-02103, No. 1 (S.D.N.Y. Mar. 13, 2007).” https://www.docketalarm.com/cases/New_York_Southern_District_Court/1–07-cv-02103/Viacom_International_Inc._et_al_v._Youtube_Inc._et_al/1/.

[16] “17 U.S. Code § 512 – Limitations on Liability Relating to Material Online | LII / Legal Information Institute.” Accessed December 14, 2014. http://www.law.cornell.edu/uscode/text/17/512.

[17] “17 U.S.C. § 512”

[18] Viacom International Inc. v. YouTube, Inc., 718 F. Supp. 2d 514 (Dist. Court, SD New York 2010).

[19] Helft, Miguel. “Judge Sides With Google in Viacom Suit Over Videos.” The New York Times, June 23, 2010, sec. Technology. http://www.nytimes.com/2010/06/24/technology/24google.html.

[20] Stempel, Jonathan. “Google, Viacom Settle Landmark YouTube Lawsuit.” Reuters. March 18, 2014. http://www.reuters.com/article/2014/03/18/us-google-viacom-lawsuit-idUSBREA2H11220140318.

[21] Helft, “Judge Sides With Google.”

[22] “Copyright Match on Vimeo.” http://vimeo.com/help/faq/legal-stuff/copyright-match#what-is-copyright-match.

[23] Witt, Darnell. “Staff Blog / Copyright Match on Vimeo.” Vimeo, May 21, 2014. http://vimeo.com/blog/post:626.

[24] “Lenz v. Universal.”

[25] Helft, “Judge Sides With Google.”