Intellectual Property
Cite as: Ari Okano, Digitized Book Search Engines and Copyright Concerns, 3 Shidler J. L. Com. & Tech. 13 (Apr. 6, 2007), at <http://www.lctjournal.washington.edu/Vol3/a013Okano.html>
© 2007 Ari Okano
Internet companies, libraries, and archives increasingly are digitizing literary information and providing access to digitized content through Internet search engines. This Article compares digital book search engines from Google, Yahoo!, Amazon.com, and MSN and highlights the different approaches to each of these models. In the fall of 2005, two copyright infringement lawsuits were filed against Google for their new search engine, Google Book Search. At issue in both lawsuits is a component of Google Book Search, Google’s Library Project, through which Google is digitizing the entire library content — including copyrighted material — of the University of Michigan library. This Article examines the limits of the fair use defense to copyright infringement in the Google cases to help establish what is permissible with respect to digitizing copyrighted materials and providing associated search features.
Introduction
Comparing Digital Book Search Engines
Digital Book Search Engines as Marketing and Educational Tools
The Google Print Lawsuits, Copyright Infringement in the U.S.?
The Fair Use Doctrine: An Affirmative Defense to Copyright Infringement
Conclusion
<1> In the fall of 2005, two copyright infringement lawsuits were brought against Google, 2 the online search giant, after Google launched an online digitized3 search engine that will eventually allow users to search over 30 million books.4 The service, Google Book Search,5 allows people from all over the world to search both copyrighted and public domain materials. Because Google’s effort is worldwide, in 2006, two foreign lawsuits were also filed against Google Book Search, one in Germany and one in France.6 The German lawsuit for a preliminary injunction has since been dropped due to an informal opinion issued by the Copyright Chamber of the Regional Court of Hamburg. This informal opinion stated that the petition for preliminary injunction was unlikely to succeed.7
<2> Google is not the first entity to digitize literary material. Non-profit efforts to digitize literary material include Project Gutenberg,8 the University of Pennsylvania’s On-line Books Page,9 and the Internet Public Library.10 In addition, Amazon.com launched its Search Inside the Books feature in 2003.11 Yahoo.com and MSN.com announced in the fall of 2005 their own efforts to create digitized book search engines and are collaborating together in the Open Content Alliance (“OCA”). The OCA is a coalition of college libraries, the Internet Archive, and the National Archive of England united to create high quality digitized copies of historical works of fiction along with specialized technical papers.12 However, Google is the most controversial digitized book search engine because the company has taken the unprecedented steps of 1) digitizing the entire collections of selected libraries (including copyrighted materials),13 and 2) requiring copyright holders and publishing houses to opt-out of Google’s Library Program if they do not want their copyrighted materials included.
<3> This Article first examines the differences among digital book search engines currently available online. Second, the Article explores the value of digital book search engines as educational and marketing tools. Finally, the Article discusses current legal implications for Google Book Search in the U.S., including, copyright infringement and whether Google may assert an affirmative defense of fair use.14 This Article examines the limits of the fair use defense to copyright infringement in the Google cases to help establish what is permissible with respect to digitizing copyrighted materials and providing associated search features.
<4> The vision of Google Book Search is to organize millions of books by putting their content where it is most easily found – in Google search results.15 Google Book Search is comprised of two separate programs, the Partner Program — which requires publishers and authors to consent before Google will digitize copyrighted material — and the Library Project.16 Through the Library Project, Google has contracted with major research libraries, called library partners, to digitize public domain books, and in some cases, copyrighted material.17 Despite the lawsuits filed against Google Book Search, many educational institutions are still contracting with Google to become Library Partners.
<5> Google sorts the materials from both the Partner Program and Library Program into various categories in order to protect copyright holder interests. If the book is in the public domain or the rights holder has given express permission through the Partner Program, Google displays a “full view” and the entire book will be displayed.18 If the copyright holder has given Google express permission through the Partner Program, but does not want the full book displayed, Google displays a “limited preview” that consists of the page on which the search term appears and a few pages before and after that page.19 If the book is still under copyright, not part of the Partner Program, and scanned through the Library Project, Google displays a “snippet view” where the viewer will see the search term and up to three snippets of text from the book showing the search term in context.20 In some instances — such as with reference books or dictionaries — there is “no preview available” and Google will display only the bibliographic information and a link to help locate the book in a bookstore or library.21
<6> Google Book Search has been subject to two lawsuits in the U.S. and a lawsuit in France, because through certain Library Partners participating in the Library Project,22 Google is digitizing copyrighted texts without prior express permission from the copyright holder.23 U.S. copyright law gives a copyright owner the ability to control reproduction, display, and distribution of a protected work.24 In order to use a copyrighted work, one typically seeks permission from a copyright owner and negotiates license terms for the use of the work.25 Google permits copyright holders who do not wish to be part of the Library Project to opt-out of the program.26 To opt-out, a copyright holder must verify that he/she holds the copyright to the material he/she wishes to exempt and provide Google with a list of books the holder does not want included in the program.27 The opt-out approach avoids the need to obtain permission from every copyright holder. It also requires copyright holders to take affirmative steps to protect their work.28 Google argues that the transaction costs of negotiating consent from all copyright holders would be prohibitive of amassing a comprehensive digitized book search engine. Whether the transaction costs of assembling a digitized book search engine of over thirty million books is truly prohibitive is an important factual issue which will play a key role in the fair use analysis of Google’s case or in any potential settlement agreements.
<7> Copyright holders assert that the natural extension of Google Book Search’s opt-out rule would lead to an overwhelming burden for copyright holders if and when others entities follow the same approach to digitizing.29 Authors and publishers fear that if the courts deem the opt-out approach as acceptable, then copyright holders would have to police established and emerging search engines for possible violations of their copyright. The creation of a clearinghouse of copyright holder’s consent for literary works is one proposed option that is meant to alleviate the supposed burden of obtaining a large number of consents. This approach has been considered since Google adopted their opt-out approach in the fall of 2005.30
<8> Amazon’s Search Inside the Books feature has an opt-in policy for all copyrighted material.31 Amazon launched the Search Inside the Books feature as a means to sell books in 2003. This feature initially allowed anyone to search the content of 120,000 books with a total of 33 million pages.32 Like Google’s limited preview and snippet policy, Amazon has limited the number of pages that a user may view to the page on which the search term(s) appear and two pages before and after that term.33
<9> Yahoo! and MSN are attempting to avoid controversy with their collaboration in the Open Content Alliance (“OCA”).34 The OCA database is comprised of only public domain materials and materials available under licensing agreements.35 Moreover, commercial publishers, such as O’Reilly Media and the University of Toronto Press, have agreed to make certain copyrighted content available to the OCA.36 Due to Amazon, Yahoo!, and MSN’s opt-in policy for obtaining copyright permission for their respective digitized copyright engines, their digitized collections are smaller than that of Google Book Search. The smaller collection sizes of Google’s competitors provided support for Google’s argument that obtaining licenses from all copyright holders is prohibitive for a company trying to develop a worldwide digital book search engine comprised of over thirty million books.
<10> Digital book search engines that access copyrighted material are beneficial to the public for several reasons. They create a point of reference to texts that enjoyed only limited success because of minimal distribution and or lack of market success. They also create reference for books that are out of print or difficult to locate. Potential buyers are able to browse an entire book for terms of interest within a digitized book collection, and locate more books than possible through a simple title search.37 With this goal in mind, Amazon, in November of 2005, announced a plan to expand the Search Inside the Books feature to allow customers to purchase individual pages of an author’s book through a new program, “Amazon Pages,” or to purchase access to the entire book online through “Amazon Upgrade.”38
<11> Google Book Search also emphasizes the marketing benefits to copyright holders, contending that the chief beneficiaries of the service are authors whose backlist, out of print and lightly marketed new titles will be introduced to countless readers.39 Publishers that participate in Google Book Search are already reporting increased backlist40 sales.41 Several of the publishers who filed suit against Google’s Library Project are at the same time participants in Google’s Partner Program. These publishers recognize the benefits of Google Book Search as a vehicle for users viewing and buying books, but object to the “massive, wholesale, and systematic copying of entire books still protected by copyrights”42 that occurs from participation in certain Library Partners in the Library Project.43
<12> Only two contracts between Google and each individual Library Partner has been made available to the public; those between the University of Michigan44 and the University of California.45 Under both contracts Google bears most of the costs of creating the digitized copies by providing both the Google employees to digitize the content, and the equipment necessary to digitize the content. Moreover, Google is providing each of the libraries with a “University Digital Copy,”46 which is to be used by the libraries for preservation purposes.47
<13> Digital book search engines arguably transcend an individual copyright holder’s interest by the public benefit derived from such engines.48 Moreover, digital book search engines create digitized copies of literary works, preserving these works against loss, damage and decay.49 Google is creating a research tool that, within ten years, plans to create an engine with over 30 million books available for users to search.50 No other digitized book search engine approaches anywhere near the number of books Google plans to digitize, which should be a significant factor in determining whether Google is entitled to assert an affirmative defense of fair use.
<14> In the fall of 2005, the Author’s Guild (“AG”) and the Association of American Publishers (“AAP”) sued Google 51 for digitizing the entire collection of the University of Michigan’s library. The focus of the lawsuits is entirely on the Library Project and its opt-out policy. The claims of copyright infringement by the AG and AAP do not include Google’s Library Partner Program, because that program is conducted pursuant to express agreements between Google and the copyright holder and thus there is no infringement. Both sets of plaintiffs seek declaratory, preliminary, and permanent injunctive relief.52 The plaintiffs in the Author’s Guild case also seek damages for infringement of copyrighted materials included in the project.53 However, both suits seek to obtain a permanent injunction against Google and to require Google to destroy its digitized copies.54 In October of 2006, Judge John Sprizzo consolidated the two cases.55 Motions for summary judgment have been delayed until January of 2008.56
<15> The key distinction between the digital book search engines is the way in which the companies address obtaining permission from copyright holders.57 At present, the digital search engines being developed and operated by Amazon, Yahoo, and MSN seem to be accepted as non-infringing uses, or as fair use. Google, Amazon, Yahoo, and MSN do not allow copyrighted materials to appear in their entirety absent express licensing agreements. Amazon, Yahoo, and MSN obtain permission from copyright holders to include their books in digitized search engines, or include only public domain materials.
<16> Copyright law was designed to protect original expression, such as literary works. A copyright holder has exclusive rights to display and perform a work publicly, to make and distribute copies of their work, and to prepare derivative copies of his or her work.58 These rights do not protect ideas, but rather arise when the work is fixed in a permanent tangible form, such as a writing or recording, and under the 1998 extension the copyright exists for 75-years after the author’s death.59
<17> To present a prima facie case of infringement, a copyright holder must show: (1) ownership of the copyrighted work in question, and (2) violation of one of the exclusive rights granted under 17 U.S.C. § 106 without the express permission of the copyright holder.60 The exclusive rights at issue are the rights to copy, display, and license the protected works. In the Google library litigation, the district court will first consider the question of whether Google’s wholesale copying of copyrighted texts in the University of Michigan’s library is prima facie infringement of the AG and AAP’s §106 right to copy. If the University of Michigan had simply contracted with Google to make digital copies for the purpose of preservation for the library, and Google had not received a copy for its own database, the act of copying itself would arguably have not been infringing under the Library Exception.61 Google, however, does receive a digitized copy and thus the library exemption does not apply.
<18> If Google’s wholesale copying is found to violate the AG and AAP’s §106 rights, the district court will consider whether Google’s opt-out procedure for copyright holders constitutes constructive permission to copy the copyrighted texts in the University of Michigan library. Google adopted the opt-out approach to the Library Project in response to the concerns of copyright holders, and provided copyright holders over two months to opt-out of the Library Project, arguably obtaining the express permission of copyright holders.62 Again, whether the district court and appellate courts find Google’s opt-out policy persuasive will depend on the strength of Google’s argument that the cost of licensing would unduly limit the scope of the digital book search engine to the detriment of society.
<19> After considering whether wholesale copying violates the AG and AAP’s § 106 right to copy, the district court will then turn to examine whether Google’s display or distribution of copyrighted materials by way of the search engine display results prima facie infringement. The snippets arguably are like quotations, and the U.S. Supreme Court has found that even substantial quotations may qualify as fair use.63 There is no specific number of words, lines, or notes that may safely be taken from a copyrighted work without permission.64 In some cases, the amount of material copied is so small (or "de minimis") that the court permits the use of copyrighted material without conducting a fair use analysis.65 To establish that the infringement of a copyright is de minimis, the alleged infringer must demonstrate that the copying of the protected material is so trivial “as to fall below the quantitative threshold of substantial similarity, which is always a required element of actionable copying.”66 The limited search results are Google’s effort to discourage users from attempting to access an entire copyrighted text online for free. Ultimately, the district court will focus on both the act of copying and the display results, thus the de minimis defense is unlikely to apply to Google and the court will engage in an analysis of fair use.
<20> Both the AG and AAP have requested preliminary injunctions on Google from digitizing copyrighted material found in the University of Michigan library. A federal court in the Second Circuit will grant a preliminary injunction where the movant can show either: “(1) irreparable harm in the absence of the injunction, and (2) either (a) a likelihood of success, or (b) sufficiently serious questions going to the merits to make them a fair ground for litigation and a balance of hardships tipping decidedly in the movant’s favor.”67 The irreparable harm requirement can be met by proof of likelihood of success on the merits, which the defendant can rebut by demonstrating that its copying is protected by the fair use doctrine.68 Judge Pierre Leval of the U.S. Court of Appeals for the Second Circuit, a noted copyright expert who has been cited by the U.S. Supreme Court,69 emphasizes that there is no statutory preference for injunctive relief, and that “the tendency toward the automatic injunction can harm the interests of plaintiff copyright owners, as well as the interests of the public and the secondary user.”70
<21> Google has raised procedural issues as grounds for dismissing the litigation, and has questioned the validity of the AG and AAP class actions.71 The AG and AAP represent copyright holders who own the copyright to a portion of the works in question at the University of Michigan Library.72 In response to the AG complaint, Google contends that some, or all, of the eight thousand authors represented by the AG are barred from asserting copyright infringement because they do not own the copyright or electronic rights to the works in question.73
<22> Assuming that the district court finds that the Library Project has resulted in either infringement in the act of copying, display, or both, the determination as to whether Google will be held liable for copyright infringement depends on Google’s ability to assert a valid affirmative defense of fair use for their alleged infringing acts. This analysis is important not simply to establish the likelihood of Google’s success, but also for examining the legal boundaries for digitizing and search features that utilize copyrighted materials.
<23> If Google is found to have infringed the AG and AAP’s copyrights by either wholesale copying or by the display of copyrighted texts, the district court will consider whether Google is entitled to a fair use defense. 74 The factors that the court will most likely emphasize are whether Google’s use of copyright materials to enable digitized searches by third parties is a “transformative use,” and whether the profit derived from advertisement revenues by Google undermines a fair use defense by depriving copyright owners of potential revenues from licensing this newly developed derivative market for displays of digitized copyrighted text.
<24> The fair use doctrine has been described as, “the most troublesome doctrine in the whole of copyright.”75 Despite the emphasis on Google’s opt-out process, fair use is not about obtaining consent, but about balancing the public interest in the use against the interests of copyright holders. Essentially, the only way to predict whether the doctrine will immunize the particular use is to analogize the facts to past fair use cases.76 In cases of new technology, like digitized book search engines, the lack of analogous cases may be problematic for parties involved in litigation.
<25> Google argues that its digitizing of copyrighted texts constitutes fair use.77 The fair use doctrine permits courts to avoid rigid application of §106 when it would stifle the creativity that the law was designed to foster.78 Fair use may encompass copying of copyrighted material done for a limited and "transformative" purpose such as criticism, commentary, parody, teaching, research, or news reporting.79 However, courts have found that non-transformative works may also qualify as fair use.80
<26> In evaluating the fair use defense there is no bright line rule distinguishing fair use from copyright infringement.81 Courts apply a balancing test between the interest in encouraging new creative works — which requires that the copyright holder maintain the ability to profit from his/her labor — and the interest in advancing knowledge through broad public access.82
<27> The test is compromised of the following four factors: (1) the purpose and character of use, where the court determines whether the use is commercial or for nonprofit educational purposes; (2) the nature of the copyrighted work; (3) the amount and substantiality of the portion of copyrighted material used in relation to the work as a whole; and (4) the effect of the infringing use upon the potential market for or value of the copyrighted work.83 In Blake v. Google, the court found that “while no one factor is dispositive, courts generally give the most weight to the first and fourth factors.”84 Leval notes that these factors are not a “score card that promises victory to the winner of the majority.”85 Leval further emphasizes that courts should “examine the issue from every pertinent corner and to ask in each case whether, and how powerfully, a finding of fair use would serve or disserve the objectives of the copyright.”86 The four factors are not exhaustive. In addition to the four factors, a court may also consider factors such as the nature of the public interest, presence of good faith/bad faith, artistic integrity, and privacy.87
<28> Several recent copyright cases provide a context for a fair use analysis of digitized search engines, Kelly v. Arriba Soft Corp., 88 Field v. Google,89 and Perfect 10 v. Google.90 These cases analyze copyright infringement with respect to digital search engines. Since the Second Circuit recently adopted the reasoning and analysis in the first case, Kelly, it will be important in analyzing the cases against Google Book Search.91
<29> In Kelly, Arriba Soft Corp. operated an Internet search engine for Internet images. Arriba created its database of pictures by copying images from other web sites and reducing the images into thumbnail images of the full size images. Arriba created this database without the express permission of the website operators.92 Like Google’s purpose, Arriba’s purpose in creating the database was to sell advertising space. The plaintiff in the case, Kelly, was a professional photographer who contracted with websites for the use of his copyrighted images.93 Kelly brought claims against Arriba for copyright infringement based on unauthorized reproduction. The lower court found that Arriba’s reproduction constituted fair use, and the Ninth Circuit affirmed.
<30> The holding in Kelly has recently been complicated by two district decisions that ruled directly on various Google search engine and display functions. First, in Field v. Google, Inc., the District Court of Nevada held that a Google caching mechanism qualifies as a fair use of copyrighted material.94 The court rationalized that because, “the search engine served different and socially important purposes in offering access to copyrighted works through cached links and did not merely supersede the objectives of the original creations, its alleged copying and distribution of the author’s copyrighted works was transformative.”95 Because of the transformative function, the court found Arriba’s copying was protected under fair use.96
<31> In the second, more recent case, Perfect 10 v. Google, Inc., the district court of the Central District of California in part granted Perfect 10’s (“P10”) request for a preliminary injunction against Google and Amazon.com. Google has filed an appeal with the U.S. Court of Appeals for the Ninth Circuit.97 P10 publishes the adult magazine “Perfect 10” and operates the subscription website, “perfect10.com,” both of which feature high quality photographs of nude models.98 P10’s complaint asserted various copyright infringement claims for copying, reproducing, distributing, publicly displaying, adapting, infringing, or vicariously contributing to infringement.99 The district court rejected the notion that in-line linking of images directly infringes a copyright owner's public display right.100 The district court also rejected the contention that Google was secondarily liable for creating the audience for the infringing website.101 However, the district court found that P10 established a likelihood of proving that Google’s creation and public display of thumbnails does directly infringe P10’s copyrights.102
<32> The finding that Google’s public display of thumbnails does infringe P10 copyrights conclusion appears inapposite to the Ninth Circuits findings in Kelly, but may be distinguishable on the basis of facts. Like Arriba, Google created thumbnail size images of P10’s high quality nude photographs.103 Unlike the photographer in Kelly, P10 licensed Fonestarz Media Limited in the United Kingdom for the worldwide sale and distribution of P10 reduced-size copyrighted images for downloads on cell phones.104 Fonestarz sold an average 6,000 images per month in the U.K.105 Thus, the facts in Perfect 10 differed from those in Kelly, because there was a derivative market for thumbnail size images that Perfect 10 had a §106 right to license.
<33> Kelly, Field, and Perfect 10 help provide a basis for determining whether Google’s digitized copies of literary materials likely qualify for the fair use defense, which is discussed below.
(1) The purpose and character of use
<34> The first factor to be addressed in a fair us analysis is the “purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes.”106 The Supreme Court asserts that most important to the court’s analysis of the first factor is the “transformative” nature of the work.107
<35> A transformative use is one that adds to or changes the copyrighted work to provide something new—whether expression, meaning, or message.108 Leval emphasizes that, “[t]he use must be productive and must employ the quoted matter in a different manner or for a different purpose from the original.”109 Google asserts that its use of copyrighted books obtained through the Library Project is transformative because its search engine allows people to find things in books via a complex search protocol program, which is different from reading the book.110 If someone types in “Madame Defarge and knitting”, for example, he will receive Charles Dickens’s A Tale of Two Cities as well as several literary criticism novels analyzing how a common woman has knit together the threads of the French Revolution.
<36> A search of the Library Project material produces text exactly as the author created it, but the difference is that the text has been catalogued in an electronic medium within a copyrighted search engine that is likened to an enhanced version of a library card catalogue. In Kelly, the Court of Appeals for the Ninth Circuit found that the use of Kelly’s images to create thumbnails searchable on Arriba’s search engine was transformative because (1) the thumbnails were smaller and lower in resolution, and (2) served a different function than the originals.111 Like the low-resolution thumbnails, Google’s digitization of library texts produces low-resolution pages of the original book, making some fine print difficult to read.112 In Kelly, the original function of the images was to produce an aesthetic experience, whereas the use of the thumbnails within the search engine was to “…improve access to images on the internet and their related web sites.” Google Book Search does not supplant the need for the original. The value of Google Book Search is the creation of a searchable index that exceeds the Dewey decimal system.113
<37> In Field, the district court emphasized that Google serves “different and socially important purposes in offering access to copyrighted works through ‘cached’ links and does not merely supersede the objectives of the original creations.”114 As in Field, the Library Project serves a socially important purpose and does not merely supersede the need for print books.
<38> The purpose of allowing transformative use of copyrighted works is to promote the sciences and arts, which in turn promote creativity and learning.115 Google Book Search serves the public interest by providing limited access to and preservation of millions of books while protecting right holders’ interests by limiting the amount of text displayed.
<39> Google’s Library Project also serves commercial goals, which include the production of advertising revenue. When evaluating the purpose and character of use, one must consider whether the use is commercial.116 The Library Project is commercial because Google gains a direct economic benefit from Google Book Search by selling advertising space. Google has pledged to show no advertising next to pages of library books.117 However, a commercial purpose does not mean that use is automatically deemed infringing. The U.S. Supreme Court in Campbell v. Acuff-Rose Music, Inc., recognized that “nearly all of the illustrative uses listed in the pre-amble paragraph of §107 … are generally conducted for profit.”118
<40> As indicated by Campbell, the question of commerciality of a use is not a binary question, and if the use is commercial the court may still determine that it is fair.119 For instance in the recent Second Circuit case, Blanch v. Koons, Jeff Koons a neo-pop artist known for incorporating images from popular media and advertising, incorporated Andrea Blanch’s copyrighted photograph “Silk Sandals” taken for the fashion magazine Allure, into his visual artwork titled “Niagara.”120 The court found that the incorporation of a portion of a copyrighted photograph into an art collage was a transformative use that lessens the importance of commerciality for the purpose of determining whether the use was fair.121 The Second Circuit held that copyright law’s goal of promoting the progress of science and art would be better served by allowing Koon’s fair use of Blanch’s work then preventing it.122
<41> In Kelly, the Ninth Circuit found that while the purpose of Arriba’s search engine was commercial, the images were neither used directly to promote the search engine nor did Arriba try to profit directly by selling Kelly’s images. Instead the images were among thousands of images in Arriba’s search database. The Ninth Circuit found the use of images was not highly exploitative and the commercial use factor weighed only slightly against Arriba in a finding of fair use.123
<42> As Kelly emphasizes, the factor of “purpose and character of use” in the Second Circuit is not merely of commercial gain, “but whether the user stands to profit from the exploitation of the copyrighted material without paying the customary price.”124 Google’s use of the digitized material is arguably less exploitative in nature than traditional types of commercial use, because users cannot access an entire copyrighted book without paying the price. Google has also taken steps to ensure that in cases where materials could easily be exploited by users to avoid paying the customary price — such as reference materials or dictionaries — no text is displayed.
<43> However, in Perfect 10, the Central District of California found that Google’s use of thumbnails was commercial, and weighed in favor of P10.125 The court distinguished Google’s search engine from Arriba’s by focusing on Google’s commercial benefit from its AdSense program.126 The AdSense program allowed Google to share advertising revenues from the infringing sites that displayed P10’s nude images and contributed to Google’s bottom line.127 It is important to note that the district court did not find that the thumbnails superseded P10’s use of the full size images.128 Moreover, the district court found that Google’s thumbnail size images interfered with P10’s licensing agreement with another company for the sale and distribution of P10’s reduced-size images to download on cell phones.129 The district court found that Google’s derivative profits from selling advertising space to infringing websites and the restated impact on P10’s right to license digitized copies to search engines tilted the commercial use factor against a finding of fair use.
<44> While commercialism may weigh against a finding of fair use, it is not conclusive. Moreover, Google Book Search does not usurp the need to pay for a book. The search engine is intended to better help users locate the appropriate book.130 The fact that Google Book Search is arguably creating transformative works with demonstrable public value may tilt the “purpose and character of use” factor in favor of a finding of fair use.
(2) The nature of the copyrighted work
<45> Creative works are the types of material intended for protection by the Copyright Act, and thus many of the books in question likely qualify as material protected by copyright.131 Whether a work is published or unpublished is also critical to analyzing the nature of the copyrighted work.132 A derivative work is more likely to fall within fair use if the original work has already been published.133 The copyrighted materials at issue are books already published and catalogued in the University of Michigan Library.134 In Kelly, the fact that the images were already displayed on the Internet resulted in the Ninth Circuit balancing this factor in favor of Arriba’s search engine.
<46> The Library Project books are already in the collection of the University of Michigan Library and can be accessed by any member of the public who goes to the library.135 However, not all of the books, if any, were previously published in digitized form. This is a factor that may weigh against Google in determining whether its actions qualify for the fair use affirmative defense.136 A court may find it significant that Kelly dealt only with copyrighted material already accessible on the Internet, whereas Google is actively transforming copyrighted material from analog form into digitized form.
(3) The amount and substantiality of the portion used in relation to the copyrighted work as a whole
<47> Courts tend to weigh in favor of the copyright owner and against the defendant claiming fair use when that defendant has copied an entire work.137 Here, Google is digitizing the book collection of the entire University of Michigan library, which should factor against a finding of fair use. However, in Kelly, the Ninth Circuit found that this factor neither weighed for or against Arriba, as it was necessary for Arriba to copy the entire image to “allow users to recognize the image and decide whether to pursue more information about the image or the originating website.”138 Like Kelly, Google must copy the entire work in order to allow users to locate books and decide whether to pursue more information about the books.139 Unlike Kelly, the full text of a copyrighted work is never displayed. If Google copied only public domain works, the user would identify fewer books with the information they seek, thereby decreasing the effectiveness of the search engine.140 Thus, the amount and substantiality of the portion used should not weigh against Google in a fair use balancing test.
(4) The effect of the use upon the potential market for or value of the copyrighted work
<48> When evaluating the effect of the use upon the potential market, the court must consider: (1) the extent of harm caused by an allegedly infringing act on a market, and (2) if widespread and unrestricted, the defendant’s infringing conduct would create a substantial adverse impact on the potential market for the original material.141 Leval argues that while the market effect is significant that the U.S. Supreme Court has overstated its significance.142 The practical concept of copyright law is that authors are rewarded for creativity, and a secondary use that interferes excessively with the author’s incentive is prohibited.143 Google recently lost a copyright infringement case in Belgium where eighteen mostly French-Language newspapers filed a complaint that Google’s cached links provided free access to articles that the paper sold on a subscription basis.144 In this case, the court emphasized the importance of secondary markets.
<49> In the U.S., courts will consider markets that are not only currently in existence, but also any potential market that a creator may reasonably develop or license.145 Depriving a copyright owner of income may violate copyright laws even if the adverse party is not competing directly with the original work.146 However, federal courts have stated that if a court concluded in every case that potential licensing revenues were “impermissibly impaired simply because the secondary user did not pay a fee for the right to engage in the use, the fourth factor would always favor the copyright holder.”147 A court will examine the impact on potential licensing revenues for “traditional, reasonable, or likely to be developed markets.”148 Thus, some loss of royalty revenue or licensing revenue may be deemed acceptable as long as a market is not impaired because the material serves the consumer as a substitute or supersedes the original.149
<50> In UMG Recordings, Inc. v. MP3.COM, Inc., the owners of copyrights to musical recordings sued MP3.COM for copyright infringement. 150 The district court found in UMG Recordings that copyright holders had the exclusive right to control derivative markets by refusing to license a copyrighted work. 151 The website MP3.COM created a derivative market by sharing or selling to “subscribers converted versions of the recordings it [MP3.COM] copied, without authorization, from plaintiff’s copyrighted CD’s.”152 The court found that the record companies had a broad right to grant or withhold a license to share or sell music, despite the fact that MP3.COM allegedly had a positive impact on the sales of musical recordings.153
<51> In Kelly, the Ninth Circuit found that the effect of the use upon the potential market weighed in Arriba’s favor. Arriba’s search engine “guided users to Kelly websites licensed to display and sell Kelly’s images rather than away from them.”154 Book search engines like Amazon’s “Search Inside the Books” have not hindered sales of authorized copyrighted materials; instead the search engine has boosted sales.155 Google Book Search is also likely to increase sales as more users will be able to discover works relevant to their interests.156
<52> However, Perfect 10 emphasized not the effect of the thumbnails on the full-size P10 images, but on the market for P10 thumbnail size images. Publishers Random House and Harper-Collins intend to create similar digitized book search engines, which may prove problematic for Google.157 However, such licensed systems would fail to include orphaned works, which are works where the rights holder is unknown or cannot be located.158 There may be independent value in a comprehensive book search system, such as Google Book Search, which incorporates licensed and orphaned works instead of giving incomplete coverage.
<53> Google’s Library Project also does not promote third party infringement of copyright, which is another concern when courts examine the effect of the copyright infringement upon a potential market. In Metro-Goldwyn Mayer v. Grokster, the United States Supreme Court examined the peer-to-peer networks of Grokster and StreamCast. 159 The court found that the software of Grokster and Streamcast was intentionally used to enable users to reproduce and distribute copyrighted music and videos without authorization. The court held that one who distributes a device with the object of promoting copyright infringement is liable for the infringement by third parties.
<54> On either Google Book Search or Amazon’s “Search Inside the Books,” a third party cannot access the entire copyrighted book, thus Google’s interface does not contribute to third party infringement. An argument raised by copyright holders is the potential risk of breach to Google’s security could endanger their works.160 Google has taken steps to protect the digitized materials, and a security breach is unlikely. The threat posed by a security breach is arguably less significant in the realm of digitized books than in pirated MP3s or DVDs because the market for e-books is slowly growing and the quality of the digitized copies made by Google has not caused large numbers of readers to switch to e-books.161 More likely, a person would attempt to assemble a print book from the Book Search Results. However, even if a person were to cut and copy part of the text that appears in a Google Book Search, under the fair use doctrine a person may use a portion of the text for purposes such as criticism, comment, news reporting, teaching, scholarship, or research.162
<55> In the Grokster case, while the court ultimately found Grokster and Streamcast liable, the court discussed the valuable benefit to society of efficient, secure, and cost effective peer-to-peer networks.163 In Sony Corp. of America v. Universal City Studios, Inc., the U.S. Supreme Court also emphasized the helpful nature of videotape recorders to society.164 In Sony, the U.S. Supreme Court found that the sale of videotape recorders (VCRs) did not constitute contributory infringement of television program copyrights. Arguably where valuable technology does not promote contributory infringement by third parties, the courts are more likely to find fair use.
<56> P rofessor Wendy Gordon, a noted copyright scholar, has concluded that: “[a]n economic and structural analysis of the fair use doctrine and its place in the copyright scheme reveals that fair use is ordinarily granted when the market cannot be relied upon to allow socially desirable access to, and use of, copyrighted works.”165 The copyright holders seek to deny citizens worldwide access to a service that benefits advances in research in the Sciences and Arts, thus given market failure, the court should find for Google.
<57> Ultimately, when balancing the fair use factors, Google’s use of copyrighted works on the Google Book Search index should result in a finding of fair use of copyrighted materials, because (1) the purpose and character of use is transformative, (2) copying the entire book is necessary for the search engine to function as in Kelly, and (3) Google Book Search does not provide derivative markets, or allow third party users to infringe on copyrights negatively affecting the literary market. However, the case law in the area of search engines offers conflicting precedent, and the outcome will depend on a balancing of the facts, the copyright owners §106 rights, and the benefits of the public’s greater access to information.
<58> Google may opt to settle the two cases and develop licensing agreements with copyright holders.166 However, a settlement would preclude the development of precedents that would assist other parties interested in developing digital book search engines. If Google sets licensing fees with copyright holders, it may price other potential digitized book search engines out of the market if they are unable to pay copyright holders the same fees as those negotiated by Google.167
<59> Whether Google settles or the court reaches a decision, the outcome of the Google Book Search lawsuits will affect other efforts to digitize copyrighted material and display digitized material on the Internet, and influence the outcome of foreign cases. Google’s actions in creating Google Book Search have caused much debate among copyright experts, publishers, authors, and the general public.168 The Google lawsuits highlight the ambiguities in the fair use affirmative defense in the context of digital technology. What is at stake is not simply an individual’s right to make and display copies of their own work, but the boundaries the government will place on efforts to make valuable information accessible to the public in digital form.
<60> What can be extrapolated from analyzing the various digitized book search engines is that the safest course is for search engine companies to always obtain permission from a copyright owner before using copyrighted material. When it is impracticable to obtain permission, use of copyrighted material should be avoided unless the doctrine of “fair use” would apply to the situation. The full text of a copyrighted material should never be displayed online unless by an express licensing agreement with the copyright holder.