Guest Blog Post by Tom Lipinski*
As I was preparing the readings for my doctoral seminar in Information Policy class the other week I ran across a Congressional Budget Office report from 2004 [Copyright Issues in Digital Media (A Congressional Budget Office Paper)]. The last part of the report evaluates four courses of action: Forbearance (doing nothing), increase the use of compulsory licensing, revise the law in favor of copyright holders, or revise the law in favor of users. Fast-forward to more recent discussion [Copyright Policy, Creativity, and Innovation in the Digital Economy (The Department of Commerce Internet Policy Task Force), July, 2013) and Orphan Works and Mass Digitization: A Report of the Register of Copyrights, June, 2015] and the same time-worn discussion of changes in the copyright law occur without a viable solution.
Discussion of all four options continues; policy makers remain standing still. Thank goodness the wheels of justice, while at times slow moving, still turn. The courts have been addressing new technologies and copyright, often through the framework of fair use.
Reflecting on the intent of fair use we see its flexibility as well as predictability. Pamela Samuelson, observes in her law review entitled Unbundling Fair Uses [77 FORDHAM LAW REVIEW 2537 (2009)]: “The copyright fair use case law is more coherent and more predictable than many commentators seem to believe. Fair use cases tend to fall into common patterns…policy-relevant clusters …promoting freedom of speech and of expression, the ongoing progress of authorship, learning, access to information, truth telling or truth seeking, competition, technological innovation, and the privacy and autonomy interests of users. If one analyzes putative fair uses in light of cases previously decided in the same policy cluster, it is generally possible to predict whether a use is likely to be fair or unfair.” She is correct in that fair use cases do point the way forward. And while everyone is talking about the Authors Guild and Google litigations, other cases are reinforcing the public benefit and fair use of efforts to help Internet users corral and harness the vast amounts of information available. These cases are relevant to us in the library world where a prime function of our service mission is to help our patrons access the ever deepening sea of information through use of technological tools such as automated indexing, search and retrieval systems, data mining, etc. These tools of function are at the core of information work.
One recent case from the summer of 2015 underscores these fair uses and differentiates lawful from unlawful uses of the documents or resources these tools locate, identify and tag or retrieve. The case is Fox News Network, LLC TVEyes, Inc., 2015 WL 5025274 (S.D. N.Y.). TVEyes, Inc. provides a service that allows its subscribers to track the usage of words or phrases of interest that appear in new broadcasts (actually the transcripts via closed captioning of such broadcasts). The service is used by government agencies, law enforcement, businesses, NGOs, members of Congress, etc. It is not available to the general public. The 22,000 plus users of TVEyes search news broadcasts for a variety of reasons: “Government bodies use it to monitor the accuracy of facts reported…. Political campaigns use it to monitor political advertising and appearances of candidates….. Financial firms… public statements made by their employees for regulatory compliance…. White House …to evaluate news stories and give feedback to the press corps…. United States Army uses TVEyes to track media coverage of military operations in remote locations…. Elected officials use TVEyes to confirm the accuracy of information reported… TVEyes provides substantial benefit to the public.”
A previous decision in Fox News Network, LLC TVEyes, Inc., 43 F.Supp.3d 379 (S.D.N.Y. 2014) concluded that “TVEyes’ copying of Fox News’ broadcast content for indexing and clipping services to its subscribers constitutes fair use.” The court found that the service “provides substantial benefit to the public” by allowing a user to “easily and efficiently text-search … 27,000 hours of programming… most of which is not available online or anywhere else, to track and discover information.” So this is yet another case in which a commercial entity that mined, stored and made retrievable an archive of content protected by copyright was found to be a fair use. Libraries can and do create such tools!
The fact-specific context of any fair use case is crucial. It is important to note that like other litigated mass digitization efforts that were determined to be fair use these scenarios do not involve simply making available large amounts of protected content for any and all seekers; rather snippets by Google, Inc., page numbers by the HathiTrust and here, new clips and not the entire news broadcast. Further the clipping service was available to a confined audience (albeit 22,000 plus subscribers) for a specific purpose other than as the news. TVEyes is not a competing news service but is used for specific purposes such as tracking frequency and trending of stories (news about the news), error and accuracy reporting and compliance monitoring.
Further, and like the Google Books case, it was unlikely that a subscriber would “go through the trouble of running countless searches to reconstruct a full broadcast.” Remember in the Google Books case, its search/retrieval algorithm and blackout patterns rendered that result impossible. Though possible here the judge found it too improbable to impact the legal analysis.
Fair use triumphs again!
The remaining four services TVEyes, Inc. provides its customers was the subject of concern in the 2015 decision: archiving by users with the contents stored solely on the TVEyes, Inc. server, sharing via email a URL link to a video clip, downloading and permanent storage of video clips on customer devices and a date-time search feature.
Content remains searchable and available on the TVEyes server for a period 32 days. A subscriber has the ability—once a clip has been identified within that 32 day window— to “archive” the clip, that is tag it so that repeat searches would not be needed to locate the clip again. The clip remains available to the subscriber indefinitely. However, the archived clips are not stored on the subscriber’s computer but remain on TVEyes’ server, but are easily retrievable by the subscriber. The court observed that the ability to archive clips for reviewing at a later date helped promote the public discourse necessary for democracy and the free exchange of ideas. As the archive function complements the service’s main search and index function, the court concluded it was fair use.
The remaining three features did not fare so well in terms of fair use. Subscribers can also share clips by sending a URL link. “The link is public, meaning the recipient does not need to possess TVEyes login credentials in order to access the video.” While this feature can further help users engage in commentary, criticism and other fair uses listed in the fair use statute the court observed the “substantial potential for abuse.” There is no control over who could eventually access the link-clip or to what use that clip would be put. However the court was quick to add that once TVEyes, Inc. develops “reasonable and adequate protections” and demonstrates the effectiveness of those measures this feature might be considered fair use.
The court concluded that allowing subscribers to download and store clips on their own devices was not fair use as the feature “goes well beyond TVEyes’ transformative services of searching and indexing.” While the ability to “download unlimited clips to keep forever and distribute freely may be attractive” the court felt it insufficiently related to the transformative public benefit the service provides. A final service provided was the ability to search by date-time. Date-time constitute about 5.5% of all searches on the service. TVEyes argued that date-time is a necessary companion to its keyword searching, i.e., clips are located by searching the closed-captioning associated with a particular clip. Closed-captioning is riddled with errors so time-date increases the chance the user will locate a relevant clip, but this is true if the user knows the exact date-time (up to a 10 minute window) of the original broadcast airing. The court however viewed the feature as a “content delivery tool for users who already know what they seek” and concluded the feature was not fair use. Stay Tuned: in late 2015 the judge cleared the way for the parties’ ability to appeal the district court to the United States Court of Appeals for the Second Circuit, the same circuit that decided the Google Books and HathiTrust cases. The fair use determinations should be upheld on appeal.
Building a library of clips might remain prohibited but assuming the URL that subscribers share via email is to a lawfully made copy of the broadcast, sharing a URL to a non-infringing source should also be lawful. The URL itself is not protected by copyright, same as a street address is not. The date-time search feature might meet a similar fate as subscribers arguably use the feature to verify what was reported on a particular date. The court even used the example of Ted Cruz’s staff not being able to locate a particular interview even though staffers knew the date and broadcast of the interview that the closed captioning mechanism translated as “ted crews.” Like the other search feature only “segments” are locatable. If this is the case the feature should be fair use as well. It is similar to the function of cite-checking that many researchers use Google Books to accomplish.
*Note: Guest blogger Tom Lipinski is Dean and Professor at the School of Information Studies, University of Wisconsin—Milwaukee