By: Drew Apperson
The U.S. Copyright Office recently provided notice that it is considering a significant antipiracy step—the development of “technical measures” (i.e., content-recognition software) that will identify or protect copyrighted works online.[1] These measures have traditionally been implemented via private collaborations among willing stakeholders, such as online service providers and rightsholders. The Office’s plan, however, is to address major issues with the existing measures, primarily inclusivity—which stakeholders must collaborate on the front-end development.[2] The Office explained that existing measures’ “strictly voluntary nature presents inherent limitations. The absence of comprehensive coverage and the exclusion of certain stakeholder interests during the development stages could hinder a measure’s sustainable success.” [3]
One of the Office’s examples of existing measures is YouTube’s Content ID,[4] which blocks millions of allegedly infringing videos each year.[5] Content ID is a “matching system that automatically identifies content that may be infringing . . . . [by scanning it] against a database of files submitted by copyright owners.”[6] The measure automatically blocks matched content from being shared on the site, notifies the uploader of their alleged infringement, and provides several options to the rightsholder: (1) block the content outright, (2) permit the content and track its viewer statistics, or (3) permit the content and receive royalties from advertisements displayed to the content’s viewers.[7] The exact amount of content that Content ID requires for a match is unknown, but possibly a single image or just a few seconds of audio may suffice.[8]
Although Content ID’s match precision and threshold may be beneficial for existing rightsholders in recognizing potentially infringing works, Content ID is known to be so effective that some argue the measure deters, if not prevents, fair use by potential rightsholders, or content creators.[9] One of the Office’s considerations is how this affects small, lesser-represented content creators.[10] The Office addressed the Ninth Circuit’s decision in Lenz v. Universal Music Corp,[11] where a homemade video of a small child dancing was automatically flagged because its background music matched a song in Content ID’s database.[12] The Office suggested that Content ID’s precision and threshold, as well as its lack of human input, resulted from insufficient stakeholder collaboration during front-end development.[13]
The Office noted that a “key feature of any future voluntary measure should . . . involve cooperation among rightsholder organizations, all sizes of [online service providers], individual creators, and users.”[14] The Copyright Office is therefore hosting consultations consisting of one plenary session in February, followed by a series of “smaller, industry-sector specific sessions” throughout the Spring.[15] The Office has provided questions for public comment, such as whether there is even a role for the government to play, and if so, how to accomplish it. Several questions are explicitly intended for specific stakeholders: rightsholders, online service providers, or users of those online services.[16] The questions are fairly broad and encourage widespread participation, and for good reason – there are a lot of stakeholders. YouTube alone provides localized versions of its site to over 100 countries[17] and accounts for an estimated 2.2 billion users worldwide, including approximately 204 million in the U.S.[18] Accordingly, “the Office also emphasized the importance of flexibility, accountability, and comprehensive reporting.”[19]
By Summer 2022, the Copyright Office should have a clearer prediction of whether it can address the extensive concerns and develop a successful, comprehensive measure.
[1] Technical Measures: Public Consultations, 86 Fed. Reg. 72,638 (Dec. 22, 2021).
[2] Id. at 72,639.
[3] Id.
[4] Id.
[5] See Number of Videos Removed from YouTube Worldwide from 4th Quarter 2017 to 3rd Quarter 2021, Statista (Dec. 7, 2021), https://www.statista.com/statistics/1132890/number-removed-youtube-videos-worldwide.
[6] Overview of Copyright Management Tools: Content ID, YouTube: Help Center (last visited Dec. 30, 2021), https://support.google.com/youtube/answer/9245819?hl=en&ref_topic=9282364#zippy=%2Ccontent-id.
[8] Katherine Trendacosta, Unfiltered: How YouTube’s Content ID Discourages Fair Use and Dictates What We See Online, Electronic Frontier Foundation (Dec. 10, 2020), https://www.eff.org/wp/unfiltered-how-youtubes-content-id-discourages-fair-use-and-dictates-what-we-see-online.
[9] Id.
[10] U.S. Copyright Office, Section 512 of Title 17, 150–52 (2020), https://www.copyright.gov/policy/section512/section-512-full-report.pdf.
[11] Id.
[12] See Lenz v. Universal Music Corp., 801 F.3d 1126, 1154 (9th Cir. 2015).
[13] U.S. Copyright Office, Section 512 of Title 17 27–47 (2020), https://www.copyright.gov/policy/section512/section-512-full-report.pdf.
[14] Technical Measures: Public Consultations, 86 Fed. Reg. 72,638, 72,639.
[15] Id.
[16] Id. at 72,640.
[17] YouTube for Press, YouTube (last visited Dec. 30, 2021), https://blog.youtube/press.
[18] Number of YouTube Viewers Worldwide, Statista (Aug. 31, 2021), https://www.statista.com/statistics/805656/number-youtube-viewers-worldwide; Online Video Entertainment, Statista (last visited Dec. 30, 2021), https://www.statista.com/markets/424/topic/542/online-video-entertainment/#overview; YouTube Users in the United States, Statista (July 20, 2021), https://www.statista.com/forecasts/1147203/youtube-users-in-the-united-states.
[19] Technical Measures: Public Consultations, 86 Fed. Reg. 72,638, 72,639.
