(From left: Dr. Andrew Rens, Prof. Rebecca Tushnet, Dr. Aram Sinnreich, Prof. Hillary Brill, Prof. Christine Farley, Prof. Kathy Kleiman, and Dr. Patricia Aufderheide, who moderated the panel)
Top level domains, algorithmic copyright enforcement, software licenses in hybrid goods and anti-circumvention laws are all being weaponized to make expansive intellectual property claims. Internet governance scholars revealed these hidden levers of intellectual property control during a panel convened by the Internet Governance Lab at the Global Congress on Intellectual Property and the Public Interest held at the Washington College of Law on Friday, September 28, 2018. The panel discussed¹ the ways in which technological and privatized processes, such as domain name disputes, rather than formal legal processes, decide the outcome of claims to intellectual property. While these processes and related technologies have received some attention, the panel was an opportunity for scholars to compare different aspects of an emerging phenomenon.
Kathy Kleiman, Visiting Fellow at the Center for Media & Social Impact (CMSI), discussed how trademarks are being used to control Internet domains with Professor Christine Farley, co-director of the Program on Information Justice and Intellectual Property at Washington College of Law which organized the global congress, and Professor Patricia Aufderheide, Founder & Senior Research Fellow at CMSI at American University’s School of Communication, who also chaired the panel.
Ms. Kleiman described how 20 years ago, she noticed problems with trademarks and domain names. Some people were registering domain names that were clearly famous trademarks, such as americanairlines.com, and some were registering ordinary words as domain names and nicknames, such a “pokey.org” and “time.com.” But large trademark owners began to claim “ownership” of dictionary words and common names and sought the right to take domain names away from ordinary registrants (together with the webpages, email addresses and listservs associated with such domain names). Time Warner, for example, claimed extensive rights over words such as time, life, people, and fortune, all based on their ownership of magazines of these names.
When the Internet Corporation for Assigned Names and Numbers (ICANN) was founded as a private company in 1998 to manage and oversee the domain name system (DNS), it gained control over one of the few chokepoints of the global distributed Internet, namely, the routing master table of the Internet that maps domain names to Internet Protocol address. The Intellectual Property Community recognized this chokehold quickly and came to the young ICANN seeking massive protections for trademarks in the DNS. ICANN’s young Noncommercial Users Constituency fought against many proposals claiming that no trademark gives a trademark owner “ownership” of a dictionary word or common name, and their rights should only extend to preventing others from using that word in connection with domain names marketing confusingly-similar goods and services. Thus, ICANN’s Uniform Dispute Resolution Policy (UDRP) was born in 1999, and the ability of a trademark owner to yank a domain name from a registrant such as a noncommercial organization, entrepreneur or individual was born. This is dispute policy, which by its very nature stretches trademarks far beyond the national borders of their registrations and far beyond the limited categories of goods and services for which they were issued.
Professor Farley described how the trademark problem takes on new dimensions at the “top level” (the level of .COM, .ORG and .NET). In 2012, ICANN began the process of creating more than 1000 new “generic top-level domains” (gTLDs), which came to include .NINJA, .LAWYER and .LOVE. Here large companies tried to exert a new level of control over “generic words,” the basic descriptive term of their business or industry. Google applied for .SEARCH as a “Closed Generic,” Amazon for .BOOK and L’Oréal for .BEAUTY (among many others) – who wanted to own both the “top level,” as the gTLD registry, and also every second level domain name as well. It was an extraordinary proposition that a registry would be the registrant of every domain name in a generic word common to entire industries – and it created an outcry across the ICANN Community, and among small booksellers, book publishers, and public interest groups around the world. While the idea was beaten back in 2014, it has risen again as an idea for upcoming rounds of gTLDs – a powerful way to control an industry by one of its largest competitors — through the DNS.
Rebecca Tushnet, Frank Stanton Professor of First Amendment Law at Harvard Law School, touched on some problems of ICANN’s handling of domain name disputes, including due process issues. She noted that the high rate of non-response (“defaults”) by registrants in domain name dispute challenges is being deemed – falsely – as an indicator of guilt. She also noted that while ICANN created systems to protect trademark owners, such as the Trademark Clearinghouse, those systems were designed to be open and transparent and instead became secret.
Ms. Tushnet then moved onto related Internet and copyright issues with the rising new danger of DMCA+ and pre-filtering of content, before it is even posted. This goes far beyond “DMCA Classic,” with the traditional notice and takedown approach. Particularly worrisome is the fact that large platforms like YouTube are now using the DMCA+ model. Here, Ms. Tushnet worried aloud whether Amazon Cloud, for example, if it shifted to a DMCA+ model, would be able to override legal fair use protections through their own private, and technologically-enforced, copyright rules. This raises concerns about other types of private platforms, such as iTunes, and not so private platforms, such as gTLDs, and their ability to impose similar rules – without the constraints of fair use and due process protections. Algorithmic systems are not actually very good at determining whether something is in copyright, and if so, whose copyright it falls under. Some actors claim to own works that are in the public domain and use automatic processes to prevent use of those works. For example, stock photograph sellers use automated systems to claim copyright over photographs from NASA, although all NASA photographs are in the public domain. While the domain name system is weaponized to make expansive trade mark claims, algorithmic copyright enforcement is similarly weaponized to claim works that are in the public domain.
Professor Aram Sinnriech, faculty fellow at the Internet Governance Lab and research fellow at CMSI, warned about algorithmic copyright enforcement. In the United States algorithmic copyright enforcement is a self-monitored practice, while the new European Union Copyright Direct actually requires it. ² Algorithmic copyright enforcement impinges on fair use, and in the United States it infringes on freedom of speech. Algorithms reduce cultural judgments to 1’s and 0’s, but copyright law must respond to fluid and changing cultural values in which context matters. Computers are fundamentally incapable of making the kind of nuanced judgments necessary. An algorithmic filtering system is police, jury, judge and executioner. This technical procedure undermines separation of powers and sets us up for decisions made at infrastructural level, which places corporations in a position of authority vis-a-vis states. By outsourcing the definition of original and copy – which is essential to Western notions of meaning, authorship, and individuality – to algorithms, we are outsourcing decisions (who gets to speak and who controls who speaks) to opaque corporate entities whose raison d’être is profit.
Professor Hillary Brill, Practitioner-in-Residence at the Glushko-Samuelson Intellectual Property Law Clinic at Washington College of Law, asked whether software licenses will end first sale. Copyright exceptions and limitations are well established in United States common and statutory law. Technological advances pose a threat to these exceptions, specifically for copyright exhaustion, also known as “first sale” in U.S. copyright law. The first sale doctrine – or the ability to resell, transfer, loan, or give away a copyrighted work — was first established 100 years ago in U.S. common law. Since then statutory provisions and judicial decisions have strengthened and broadened the doctrine.
The first sale doctrine for physical goods, such as books, has never been stronger than it is now in the United States, which today has the broadest first sale protection possible: international exhaustion. However, the advent of “smart” technology in the form of software-embedded devices undermines this protection as it only applies to so-called hard goods. New products that include software-embedded devices no longer fit that traditional paradigm and therefore first sale protections that have allowed for free trade of copyrighted materials after a first sale have been basically gutted. It is a hidden, relatively unknown intellectual property lever that intellectual property can be used to secretly undermine and challenge previously unfettered trade of goods.
Dr. Andrew Rens, a postdoctoral research fellow at the Internet Governance Lab spoke about anti-circumvention and the Internet of Things. Anti-circumvention rules were created to prevent people copying digitally encoded music and movies. The World Intellectual Property Organization Copyright Treaty, and the United State Digital Millennium Copyright Act aimed to preserve markets for the sale of artefacts encoding sound recordings and cinematographic works. But while the markets for music and movies have changed completely the criminalization of changing software has remained static. As software is incorporated into more and more products, anti-circumvention rules are preventing people from securing the things they own against hacking and corporate surveillance. Neither established Internet governance institutions nor the World Intellectual Property Organization (WIPO) seem equipped to respond to these unintended outcomes of anti-circumvention rules. Software licenses, technical protection measures, and anti-circumvention laws together wrest control from owners, and give it to actors making aggressive copyright claims.
ICANN domain name processes, software licenses in smart things, algorithmic filtering and take down, and technical protection coupled with anti-circumvention all enable far greater control of what is claimed to be intellectual property than the rights granted by trademark and copyright. Yet these levers of control are hidden from scrutiny. Because they rely on technology and technical processes rather than legal enforcement processes the do not have safeguards built into legal processes.
1. The panel followed the general rule at the Global Congress which is that each panelist expressed her personal opinions and cannot be regarded as speaking on behalf of any organization with which she is affiliated.
2. While the Directive does not actually use the words upload filter or algorithmic enforcement there is no other practical way for most service providers to comply with the directive, so it effectively requires algorithmic copyright enforcement.