Copyright bots, otherwise known as content recognition software, are automated programs that can analyze audio and video clips uploaded to a platform, then compare those Clips against a database of content provided by copyright owners to identify matches. The copyright owners can then review the identified matches to assess if they are actually copies, if they are authorized or not, and if any action is warranted. Some programs utilizing copyright bots offer their own enforcement procedures to customers, while other programs are partnered with law firms that will act on behalf of the copyright owners to enforce their rights, including sending demand letters and filing lawsuits.
The benefits of copyright bots for copyright owners are significant and groundbreaking. Historically, to identify unauthorized uses of their works, copyright owners have had to rely on their own manual searching or notifications from third parties. With copyright bots, copyright owners can easily and more comprehensively identify unauthorized uses of their works, enforce their rights, and expand their ability to monetize their works, all in a relatively effortless and sweeping manner.
Such benefits, however, can lead to potential abuses.
First, content recognition software is not perfect. The software relies on algorithms to compare “fingerprints” and assess if a targeted clip matches the tracked clip in the software’s database within a certain range of probability. Copyright owners can often adjust the level of similarity required between clips to find a match, meaning the software may not always identify an exact copy. Considering the widespread use of sampling in music, it is not inconceivable that a copyright bot could identify a match between two audio clips, yet the copyright owner of the song being tracked in the bot’s database may not even own the portion of the song shared between the clips.
Second, copyright bots disproportionately favor the interests of copyright owners. Many programs utilizing copyright bots receive commissions on any licensing fees or settlement payments collected by the copyright owner from matches identified by the bots, such that the programs are incentivized to flag as many close matches as possible, even if they are not identical. Moreover, there is little disincentive for copyright owners to submit claims against content identified by the copyright bots, even if the match is not identical. In contrast, there may be more substantial risks for the uploaders to contest the claims, even if for valid reasons, especially if not well-versed in copyright law; such risks could include the copyright owner demanding a higher licensing fee or settlement payment, the platform blocking your account, or being sued for infringement and forced to defend.
For example, NYU School of Law recently found itself caught in a tricky catch-22 on YouTube due in part to copyright bots. Last year, NYU Law posted a video to its YouTube page of a panel featuring the musicologist experts for the opposing parties in the “Blurred Lines” copyright infringement case. In the panel, the experts played short clips of the songs in dispute to demonstrate how they analyze and explain music in copyright litigation. Once the video was posted, it was promptly flagged by YouTube’s Content ID system with multiple claims of unauthorized use of the songs. NYU disputed the claims, arguing that the use of the songs as part of a legal discussion about the analysis of music in copyright litigation was fair use, but the copyright owners maintained the claims. Having multiple claims against the video, NYU faced a dilemma—appeal the denial of the disputes to YouTube to keep the video posted, feeling confident that the use qualifies as fair use, but risk receiving multiple copyright strikes and possibly losing NYU Law’s YouTube account, or don’t appeal the denial and concede to the claims. Fortunately for NYU Law, the claims were eventually withdrawn (though not through YouTube’s process). Most other YouTubers in a similar situation without an expertise in the law of copyright and technology or the same financial stature probably would have conceded to the claims to avoid the risk of losing their accounts completely.
While copyright bots can ease the burden for copyright owners to police and monetize their works, the potential inaccuracies of the bots and the systems that operate them may lead to less than desirable results. In one reported instance, multiple claims were made against a 10-hour video of white noise uploaded to YouTube, even against sounds that the uploader had created himself. In another reported case, a pianist received a claim from Sony Music for a video he posted performing a song by Bach, a composer who is long deceased and whose works are clearly within the public domain. Part of the problem is that copyright owners often rely on the bot findings without adequately reviewing and analyzing each match to consider if it could constitute a fair use or if it even copies the owner’s own content.
Regardless, if you receive a claim or demand alleging copyright infringement, even if it seems absurd, it should be taken seriously and not ignored. Many copyright owners are aggressive in policing their works and may initiate suit for infringement if you do not respond to their demands. You may have counter-arguments and defenses against the claims, such that you may not be powerless before the copyright bots, but it is best to seek legal counsel to review and analyze the specific facts to determine how to respond to such claims, potential settlement strategies and amounts, and to what extent to fight back.