It’s best to start at the beginning.
What the DMCA is
I won’t go into all the details of the DMCA, just the bits pertinent to this discussion. The Digital Millennium Copyright Act of 1998 is a United States law that was the government’s best attempt to update copyright protections for the digital age. Heavily influenced by industry lobbyists, the bill strengthens and lengthens copyright protections and makes it illegal to get around measures designed to protect copyrighted works, like Digital Rights Management, more commonly known as “DRM”. The DMCA also has measures to stop piracy and distribution of copyrighted materials, which is what concerns YouTube.
A Historical Perspective
In internet technology, 17 years is an eternity. Apple finally gave up on its Newton product line and introduced its first iMac in 1998. Windows 98 was brand new. Microsoft stopped supporting it in 2006. In 1998, Internet Explorer nudged out Netscape as the dominant web browser. Netscape died in 2003, and Internet Explorer will be taken off life support when Windows 10 is released this year. Sony released its 3.5″ High Capacity Floppy Disk in 1998, but it could not unseat Iomega’s Zip Disk, which was already losing ground to CD-R technology, which has been supplanted by DVD-R technology. In 1998, only 13% of all U.S. households had internet access, the vast majority of which were connecting with 56 kbps modems. In 1998, Napster (which flamed out in 2002) was still a year away. The law is as antiquated as any of those technologies. Lawmakers of the time struggled to grasp the new landscape of digital copyright and tried to anticipate future technologies and their issues as well. Google filed for incorporation September 4 of 1998, and the DMCA was made law not even two months later on October 28. YouTube wouldn’t exist until 2005 and it was bought by Google in 2006.
The Lawsuit That Changed YouTube
Media giant Viacom International used the DMCA to sue YouTube. Viacom accused YouTube of copyright infringement on a large scale by making available Viacom’s own copyrighted content – shows from networks like Nickelodeon, MTV, and Comedy Central – seeking a $1 billion settlement. YouTube used the “Safe Harbour” defense, claiming that they only indirectly violated copyright, since it was the users, not YouTube itself that uploaded the content. The bitter court battle lasted over three years before being settled with no damages awarded. During those three years, YouTube scrambled to appease Viacom by finding ways to address its accountability issue. “Safe Harbour” limits the liability of Internet Service Providers (ISP) and search engines for copyright infringement, however, the ISP or search engine must respond “expeditiously” (quickly) to an allegation of copyright infringement by removing the content. This is done through a DMCA Takedown Notice. The validity of the allegation would be determined in a court of law, but thanks to Safe Harbour, if your goal is to remove the content from the internet, that happens by just making the allegation. Most disputes end with the takedown.
How the DMCA affects YouTube
Viacom’s lawsuit opened the floodgates and a rush of new court cases started piling up, raising the stakes to literally life-or-death for YouTube. Google’s first problem they had to solve was how to quickly and efficiently identify copyright infringing content. Once identified, copyright claims could be more easily pursued. With 300 hours of video being uploaded every minute, Google realized it would be impossible to hire enough humans to do the job. In order to satisfy Viacom’s demand for accountability, Google needed an automated system that would tirelessly scan for and identify copyright-infringing content on its servers. Google’s Content ID Match system was born.
How YouTube’s Content ID system works
As new YouTube content is uploaded, it is automatically checked against a reference database of copyrighted works, and Content ID bots continuously crawl the massive YouTube database with updated reference profiles, searching for more matches. When Content ID finds a match, the video is flagged for action.
How the Content ID reference database is built
The exclusive copyright holder uploads their copyrighted material into YouTube’s reference database, which now contains over 3 million works. The automated submission system handles almost 1300 works every day. YouTube warns copyright owners that they must be able to prove their ownership, but as discussed earlier under the Safe Harbour provision, YouTube does not need to verify the copyright ownership themselves. If a dispute goes all the way to court and the claimant does not actually own the copyright, the claimant faces a perjury charge. The problem is, copyright disputes rarely get to court, and unscrupulous people know this.
What happens to matched content
What happens after Content ID finds a match depends on what the copyright owner specified when they uploaded their file to the reference database. In most cases, the copyright owner elects to “monetize” the video by having YouTube place an advertisement on the video, and revenue generated from ad views goes back to the copyright owner. They could also opt that the video be blocked in certain countries, or blocked completely, or the audio could be muted (if the match is on an audio file).
Problems with Content ID
The system was designed to produce “fuzzy matches”, which gives the Content ID system the ability to recognize snippets of content, or content being played in the background of an audio track, for example. This “fuzzy match” ability often produces a lot of “false positives”. Appealing a Content ID claim is a time consuming and frustrating process. Many users feel it is not worth the effort, and either remove their videos or let them stay with ads in place. Also, many copyright owners use the Content ID system to quickly and easily find copies of their copyrighted work on YouTube. This created a flood of DMCA Takedown notifications and again, Google could never hire enough staff to handle these requests quickly enough to meet the requirements of Safe Harbour. The Motion Picture Industry Associtation of America (MPIAA) alone filed almost 12 million takedown notices with Google in one six month period. Google devised its automated “DMCA Takedown” system before YouTube even existed.
YouTube’s DMCA Takedown system
Google’s Takedown system was adapted to YouTube. Since YouTube users have accounts in order to upload, the Takedown Notices could be sent directly to the users through a web form that allows copyright owners to file infringement notifications. Once the form is submitted by a copyright owner, the disputed content is immediately removed from the user’s YouTube channel while the dispute is being resolved. If too many videos are claimed to be in violation, the channel is shut down entirely.
Starting a Takedown Notice
Google needs just six pieces of information from the copyright owner in order for them to submit a Copyright Infringement Notification (aka “Takedown Notice”). Their contact information (which they don’t verify), a description of the content, the address of the allegedly infringing video, a statement that you believe this to be infringing content, an acknowledgement of the risk of perjury, and their signature, which electronically is just their name. Note that proof of ownership of copyright is not required. Once they provide this information, the Notice is sent to the YouTube user and the video is immediately taken down.
Appealing a Takedown Notice
If you receive a Takedown Notice, you can file a counter-notice. The counter-notice requires that you provide your contact information (to be given to the claimant), the video’s address, your explanation of why you feel it is not in copyright violation, a statement swearing that your counter-notice isn’t fraudulent, and your signature. Once submitted, the claimant has a set amount of time (about three weeks) to respond to YouTube, informing them that the matter will be pursued in court, and the video stays down. If the claimant doesn’t respond to YouTube, YouTube reinstates your video. The claimant can also decide to retract their claim of copyright infringement.
How YouTube’s Content ID and Takedown systems can be abused
Unscrupulous people will submit their content for inclusion in the reference database, and either upload content they do not own and let the Content ID system flag and monetize videos for them, or they may file Takedowns and back out at the counter-notification stage, messing up the YouTube user’s channel and reputation, while avoiding the trip to court and the perjury charge. Filing enough takedown Notices at once can have a channel permanently removed.
YouTube’s DMCA Measures Are Far Too Aggressive
People do abuse it to have files removed from YouTube. A 2005 academic study of Google’s DMCA Takedown claims showed that almost 40% of its Takedown Notices issued “were not valid copyright claims“. In order to avoid further Viacom-scale lawsuits, YouTube’s automated systems are biased heavily toward copyright owners and their concerns. Another way to put it is that YouTube’s systems are built on the premise that its users are guilty until proven innocent. Proving yourself innocent is quite a daunting task and requires a trip to court.
If YouTube were a responsible organization that wanted to protect its users, it would develop something like a “white list”, where users can work to build their reputation for uploading copyright-free works, and allow some level of protection against abuse of the DMCA Takedown system by predatory or malicious claimants. For instance, there are several channels dedicated to uploading only Public Domain works, for which nobody can own the copyright, but false Content ID matches and abuse of the DMCA Takedown system result in thousands of these videos disappearing and channels being shut down regularly. Unfortunately, such a protective measure would likely require a new amendment to the DMCA or another legal action, perhaps as a class-action lawsuit on behalf of YouTube users adversely affected by these Takedown abuses. If only legislation could be upgraded like all that obsolete 1998 technology.