The Columbia Journal of Law & the Arts

The Last Line of Defense: Addressing Section 512(g)’s Dwindling Capacity to Protect Educational Fair Users on the Internet

2022-02-02T17:12:32+00:00

The COVID-19 pandemic has rapidly transformed education from one of the least digitized sectors in the U.S. economy to a largely online phenomenon, with up to 93% of households with school-age children relying on distance learning. The value of online educational opportunities has extended beyond traditional purveyors of education as well, with online service providers (OSPs) like YouTube reporting an increase in average daily views for educational videos produced by subscribers (“users”).

The rise of user-generated content in online education (“educational content”) is merely part of a larger sea change as more content is uploaded to OSPs than ever before. But educational content provides a curious case study. While educational content is a nebulous concept to describe, here it may be defined as materials—from video lectures to digital course packs—that are created wholly or in part by users, uploaded to OSPs, and designed to aid in online educational pursuits. Because copyrighted works play a central role in the explication of such topics as literature, art, and music, any increase in educational content will also inevitably bring with it an increase in the use of copyrighted materials. Even though educational uses of copyrighted works are often considered to be fair use, users who incorporate these works into their educational content remain vulnerable to infringement claims, which can cause their educational content to be removed.

This lack of protection for educational content applies to all users—“non-institutional” or “institutional”—who create and upload teaching materials containing portions of copyrighted works. Educational content therefore provides a compelling case study illustrating the limited defenses afforded to all user-generated content, whether or not it is educational in nature: If fair use educational content is considered non-infringing but nevertheless subject to removal under the guise of copyright infringement, then any non-infringing user-generated content is potentially vulnerable to the same fate. Indeed, in response to the COVID-19 pandemic, YouTube increased the automation of its takedown practices, making it easier for rightsholders to have content that incorporates their copyrighted works removed—regardless of the purpose of the use.

While the situation may look bleak for producers of both educational and non-educational content, users possess a shield against erroneous takedown notices. Enacted as part of the Digital Millennium Copyright Act (DMCA), the § 512(g) counter-notification measure grants OSP users the ability to contest the takedown of their content. However, studies have shown that this feature may generally be under-utilized, both in terms of overall volume and in relation to the amount of potentially invalid takedown notices that may warrant its use. Moreover, given additional evidence that submitted counter-notifications may be erroneous more often than not, this Note argues that the main issues plaguing the counter-notification system point to an overdeterrence of the wrong users—that is, users such as university professors and other educational content creators who are permissibly using copyrighted materials.

This Note provides a more thorough examination of the issues besetting the counter-notification process, particularly with respect to its use in education. In Part I, I discuss the functionality of the counter-notification process within the greater legal and historical background of the DMCA, focusing on the unique challenges faced by educational content creators who seek to use counter-notifications on YouTube. In Part II, I discuss the tripartite problem facing counter-notification petitioners: (1) Counter-notifications are under-utilized overall; (2) counter-notifications are over-utilized by pirates for whom the educational fair use defense does not apply; and (3) the timeframe for “putting back” material is costly to users and rightsholders alike. Finally, in Part III, I propose a “First Shot” system whereby a preemptive “counter-notification” can be used to assert educational fair use and delay takedown of educational content before a rightsholder acts. The First Shot therefore substitutes for the regular counter-notification, a privilege only for educational “Fair Users.”

Beyond Whack-a-Mole: Content Protection in the Age of Platform Accountability

2022-02-02T16:46:04+00:00

This is a story about David vs. Goliath. But with a twist. What happens when the main characters change roles before the story is over? The person whom we are rooting for becomes less clear. Are they the villain or the hero? And, because it’s a story about copyright, and I now work in the film business, let’s put it in those terms: It’s realizing halfway through the movie that maybe the supposed villain, Maleficent, wasn’t so bad after all, or at least she had a very good reason for her distrust of the humans—they did cut off her wings and stuff them in a box.

I may be going a bit far with my analogies, and the description of the overall public sentiment towards the protagonists of my story is perhaps a bit hyperbolic. Indeed, many people would probably argue that there are actually no heroes in this particular narrative. But as I discuss, there has unquestionably been a disparity between the public’s perceptions of copyright creators versus technologists throughout the years. This Article assesses how the story of “Copyright and Technology” has progressed over the years and affected related policy making, how the narrative has recently shifted, and how that shift might actually provide a basis for more cooperative efforts between the two “sides” to address Internet piracy, not fewer.

Part I of this Article explores the early development of the Internet, the regulatory approach to platform responsibility at that time, and the contrasting pre-existing negative views towards copyright. Part II describes the resulting permissive legal regime, including the adoption of the Digital Millennium Copyright Right Act (DMCA), early Internet case law addressing online copyright infringement under that legislation, and the development of the Communications Decency Act (CDA). Part III highlights the ongoing rise in piracy after the DMCA and backlash to congressional attempts to address it. Part IV discusses recent regulatory and public scrutiny of online platforms and the reassessment of their responsibility for addressing illicit conduct occurring through their services.

Trademark Fame and Corpus Linguistics

2022-02-02T17:00:07+00:00

Trademark law recognizes that the same word can mean different things in different commercial contexts. Legal protection might extend to two or more owners who use the same symbol (like Delta) to indicate different sources of disparate goods or services, such as airlines and faucets. Generally, only those uses that threaten to confuse consumers—the use of similar symbols on identical or related goods—are subject to legal sanction.

But the law extends special protection to famous trademarks, not only against confusing use, but also against dilution: non-confusing use that blurs or tarnishes the distinctiveness of the famous mark. The result of protection against blurring is that the law treats the famous mark as if the sole proper use of the term in the commercial context is to designate goods and services from the famous mark’s owner.

Protection against dilution extends only to famous marks, but courts and scholars apply differing standards for assessing fame. Nonetheless, the trend over time has been to treat fame as a threshold requiring both sufficient renown—the famous mark must be a household name—and relatively singular use.

This article argues that corpus linguistic analysis can provide evidence of whether a mark is sufficiently prominent and singular to qualify for anti-dilution protection. Corpus linguistics detects language patterns and meaning from analyzing actual language use. This article draws data primarily from two large, publicly accessible databases (corpora) to investigate whether litigated trademarks are both prominent and unique. Courts and parties can consider frequency evidence to establish or refute prominence, and contextual evidence like concordance and collocation to establish relative singularity.

Corpus evidence has some advantages over standard methods of assessing fame. Corpus evidence is cheaper to generate than survey evidence but may be equally probative. Corpus analysis can help right-size dilution litigation: A litigant could estimate the prominence and singularity of an allegedly famous mark using corpus evidence prior to discovery and better predict whether the mark should qualify for anti-dilution protection. Judges should be able to rely on the results of corpus analysis with reasonable confidence. Additionally, corpus evidence can show use of a mark over time, providing courts with tools to assess when a mark first became famous, a question that a survey generated for litigation cannot readily answer.