Media and Communication (ISSN: 2183-2439) 2015, Volume 3, Issue 2, Pages 42-52 Doi: 10.17645/mac.v3i2.270 Article The Copyright Surveillance Industry Mike Zajko Department of Sociology, University of Alberta, Edmonton, T6G 2H4, Canada; E-Mail: zajko@ualberta Submitted: 9 April 2015 | In Revised Form: 20 June 2015 | Accepted: 13 July 2015 | Published: 30 September 2015 Abstract Creative works are now increasingly distributed as digital “content” through the internet, and copyright law has created powerful incentives to monitor and control these flows. This paper analyzes the surveillance industry that has emerged as a result. Copyright surveillance systems identify copyright infringement online and identify persons to hold responsi- ble for infringing acts. These practices have raised fundamental questions about the nature of identification and attrib- ution on the internet, as well as the increasing use of algorithms to make legal distinctions. New technologies have threatened the profits of some media industries through copyright infringement, but also enabled profitable forms of mass copyright surveillance and enforcement. Rather than a system of perfect control, copyright enforcement contin- ues to be selective and uneven, but its broad reach results in systemic harm and provides opportunities for exploita- tion. It is only by scrutinizing copyright surveillance practices and copyright enforcement measures that we can evalu- ate these consequences. Keywords algorithmic; copyright surveillance; copyright enforcement; identification; internet Issue This article is part of the special issue "Surveillance: Critical Analysis and Current Challenges", edited by James Schwoch (Northwestern University, USA), John Laprise (Independent Researcher) and Ivory Mills (Northwestern University, USA). © 2015 by the author; licensee Cogitatio (Lisbon, Portugal). This article is licensed under a Creative Commons Attribu- tion 4.0 International License (CC BY). 1. Introduction internet traffic? What are the consequences of these determinations for data flows as well as people? Much of our daily lives now involve interacting with In short then, my research questions are about digital “content”. The relationships we have with these identification based on digital traces. One set deals digital goods are governed in part by intellectual prop- with identifying traffic flows and content as intellectual erty rights, and a new industry has developed to take property, the second set deals with identifying people advantage of this legal fact. The copyright surveillance and holding them accountable for traffic flows. Traffic industry monitors the distribution and use of copy- and content are identified through algorithmic compar- righted works, identifies instances of copyright in- isons to known “signatures” or characteristics. Individ- fringement, and responds against allegedly infringing uals can be identified by comparing numeric identifiers uses and individuals. Dedicated companies use auto- (IP addresses) recorded by monitoring software to logs mated methods to operate at enormous scale, scan- maintained by internet service providers. Both meth- ning millions of hours of audio and video each day, and ods can result in misidentification and reduce the com- bringing suit against hundreds of thousands of individ- plexities of copyright law to opaque decisions made by uals each year. The questions I am posing are: how are automated systems. My paper concludes by analyzing the data packets and digital fragments passing through copyright trolling, a specific kind of surveillance and en- our computer networks identified as copyrighted con- forcement that combines the two forms of identifica- tent? How are these digital flows traced to identifiable tion (of content and persons) in a particularly exploita- individuals, and how are persons held responsible for tive manner. I argue that systemic harms result from Media and Communication, 2015, Volume 3, Issue 2, Pages 42-52 42 today’s wide-ranging copyright regimes, although cop- property rights”. The rightsholders of this digital con- yright’s enforcement remains contingent and uneven. tent are often part of massive industries (most notably The internet has been seen as a threat to copyright, the music and film industries) that have turned their but copyright surveillance and enforcement technolo- attention to the internet since the 1990s. As a conse- gies have also come a long way in the internet era, and quence, a new industry has developed to offer copy- are now the tools of a profitable industry. right surveillance and enforcement as a service. This is Copyright is a convoluted body of law with strange an industry that depends on the fact of infringement to consequences for a digitally-networked society. It justi- support its existence (Lobato & Thomas, 2013), even as fies constraints on our behavior, but is also routinely it ostensibly fights to stop it. violated as we go about our daily lives (Greenberg, Copyright was originally developed to regulate pub- 2014, pp. 82-83). Copyright allows people and institu- lishers and booksellers dealing in unauthorized copies, tions to claim a monopoly in the use of a piece of writ- but today all of our computers make and circulate cop- ing, or an image, or the tiniest fragment of recorded ies of cultural goods. Before home computers and the sound. It enforces scarcity by restricting copying—an internet, a great deal of copying and circulation also act that is essential to human creativity (Cohen, 2012, took place on a regular basis. People made photocop- ch. 4) and also what our computers and networks are ies, VHS and cassette recordings, sang popular lyrics, designed to do best. Because of this, digital networks and repurposed melodies. But this behavior largely es- have threatened copyright, but they also allow for per- caped the notice of copyright owners. It was ephemer- vasive forms of copyright surveillance and enforcement al, dispersed, impossible to track and difficult to con- that have become the business model of a dedicated trol. The internet is different. Its traffic is visible by industry. The sweeping scope of this industry has dis- default, and content can be accessed around the rupted the traditional “equilibrium” of copyright’s un- world. The amount of internet traffic that arguably in- der-enforcement (Balganesh, 2013b; Lessig, 2001, pp. fringes copyright has become so large that human in- 249-250). Copyright can now be enforced against per- tervention cannot possibly keep up with it. sons and actions that would have previously escaped Fortunately for copyright owners, we now live in an copyright owners’ attention, but this enforcement is age of algorithmic surveillance and algorithmic en- uneven and inconsistent. Rather than seeking total forcement (Depoorter & Walker, 2013, pp. 333-335). control over the distribution of creative works, copy- Because algorithms are poor at adjudicating the intri- right enforcement is selective, tolerating some uses cacies of copyright law, these systems regularly gener- and intervening against others. This is because copy- ate false positives (see Depoorter & Walker, 2013, p. right depends on private actors bringing forward claims 335; Katyal, 2009, pp. 414-415). But they are effective of infringement, and the pursuit of such claims is not enough to serve the interests of copyright owners, and always advantageous or desirable. On the other hand, are having a massive effect on the availability of online some of the actors described below have built busi- content. Algorithmically-selected links are now hauled nesses dedicated to pursuing “profit-based litigation” off the web by the million, and with the assistance of (DeBriyn, 2012) and demanding monetary settlements internet service providers (ISPs), thousands of allegedly from scores of alleged infringers. infringing individuals can be identified and threatened Copyright surveillance is an international business, by the agents of copyright owners (copyright surveil- and the copyright enforcement actions that follow an lance companies and law firms). identification of infringement are often carried out Below, I outline the global scale of the copyright without state involvement. But state-backed legal re- surveillance and enforcement industry and analyze gimes remain in the background, with their threats of some of its common practices. This analysis is fur- liability and sovereign violence. The internet some- thered by internal emails and documents from Medi- times still seems like a lawless place that frustrates aDefender—once one of the industry’s most notable state controls, but it consists of physical networks companies. While MediaDefender surveilled peer-to- based in territories and jurisdictions. It is these net- peer networks, algorithmic copyright surveillance and works’ territorial basis that allows the state-backed enforcement is increasingly built-in to internet ser- monopolies of copyright to have any meaningful effect. vices, with YouTube’s Content ID system being the This also makes copyright and the industries it supports most notable example. My paper discusses the case of vulnerable to legal and political reforms. a video caught by Content ID’s extra-judicial copyright enforcement system, before closing with a more recent 2. Surveilling Digital Flows for Intellectual Property trend, in which BitTorrent surveillance services have been used to drive profit-seeking copyright lawsuits in Much of the data circulating through our networks can