Google Books and Fair Use: A Tale of Two Copyrights?

Despite not being the object of this comment, it is worth examining the ruling denying the approval of the Amended Settlement Agreement (hereinafter ASA). ¹⁰

The ASA allowed Google to continue – on a non-exclusive basis – to digitize books, sell subscriptions to databases, sell online access to individual books, sell advertising on pages from books and make other uses. Rightholders could remove their books from the database or exclude specific uses. Google would split revenues with the rightholders, paying them 63% of all revenues received from these uses (and revenues were to be distributed according to an agreed plan). A Book Rights Registry was to be established to collect and distribute the revenues.

As for books digitized before 5 May 2009, Google would pay $45 million, and minimum amounts were set for its distribution ($60 for work, $15 for entire insert, $5 for partial insert).

The ASA also provided that access to the scanned books could also be available, through participating libraries as well as through institutional subscriptions for academic, corporate and government libraries and organizations.

The ASA was not approved because – according to Judge Chin – “it would simply go too far”.

Among other issues, ¹¹ copyright – including the problems raised by orphan works and out-of-print works – and antitrust were the main reasons for its denial. Judge Chin was not comfortable with the “opt-out” system set in the ASA:

[I]f copyright owners sit back and do nothing, they lose their rights. Absent class members who fail to opt out will be deemed to have released their rights even as to future infringing conduct. … It is incongruous with the purpose of the copyright laws to place the onus on copyright owners to come forward to protect their rights when Google copied their works without first seeking their permission.

Public domain works fell outside of the settlement, but orphan works and out-of-print works constituted a large part of it. The ASA granted Google a “default” right to display out-of-print books unless the rightholder expressly opposed it (again, the opt-out system). Orphan works would also be de facto left in Google’s hands (since no one would be opposing their use or claiming any revenues from them). As Judge Chin explained, “The questions of who should be entrusted with guardianship over orphan works, under what terms, and with what safeguards are matters more appropriately decided by Congress than through an agreement among private, self-interested parties.”

On the anti-trust front, the ASA would give Google “a significant advantage over competitors” and “a de facto monopoly over unclaimed works,” basically rewarding Google for being the only one engaging in massive copyright infringement (“wholesale, blatant copying, without first obtaining copyright permissions”).

For all these reasons, approval of the ASA was denied, but the judge urged the parties to negotiate a revised settlement agreement and specifically to revise the ASA from an “opt-out” settlement into an “opt-in” one.

On 14 November 2013, the District Court granted Google's motion for summary judgment ¹² and a judgment was entered in favour of Google dismissing the case on the grounds of fair use.

The judgment is based on a careful exam of the application of the fair use doctrine in the specific circumstances of the Google Books project. According to section 107 USCA:

Notwithstanding the provisions of sections 106 and 106A, the fair use of a copyrighted work, including such use by reproduction in copies or phonorecords or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright. In determining whether the use made of a work in any particular case is a fair use the factors to be considered shall include—

The Court first quoted several US Supreme Court rulings to remind us that “from the infancy of copyright protection, some opportunity for fair use of copyrighted materials has been thought necessary to fulfil copyright’s very purpose, to promote the progress of sciences and useful arts”. In this light, the four factors of the fair use doctrine were analysed – as a general guidance – and weighed together.

The court took into account the project as a whole and found that the purpose of the use was “highly transformative”. “Google Books digitizes and transforms expressive text into a comprehensive word index that helps …. find books.” To that extent, it referred to Perfect 10 ¹³ and Arriba Soft, ¹⁴ where the Ninth Circuit held the use of works as thumbnails to facilitate online searches to be “transformative”. According to the District Court, Google Books has “transformed book text into data for purposes of substantive research.… Words in books are being used in a way that they have not been used before. Google Books has created something new.” The court found this to be the key consideration of any finding of fair use.

Furthermore, the fact that a use is commercial tends to weigh against a finding of fair use (Harper & Row, ¹⁵ Campbell ¹⁶ ) but does not necessarily prevent it (Blanch, ¹⁷ Graham Archives, ¹⁸ Castle Rock). ¹⁹ The court acknowledged that Google is a for-profit entity and obtains commercial gain from the project, but this was outweighed by the fact that it does not engage in the direct commercialization of copyrighted works (“Google Books does not supersede or supplant books because it is not a tool to be used to read books”) and by considering the public interest (“important educational purposes”) of the project.

The court concluded that the first factor “strongly favours” a finding of fair use.

Highly creative works usually have stronger protection – in terms of fair use – than factual works (Steward v. Abend). ²⁰ Many books used in the project are indeed fictional, but here the court took into account the fact that “the vast majority of the books in Google Books are non-fictional” and that they all are published and available to the public (Arica, ²¹ New Era) ²² to also “favour” a finding of fair use, on the second factor.

Google incurs in verbatim copying (scanning/digitizing, indexing and storing) of the full text of the whole work. The court pointed out that copying the entirety of a work might still be fair if it was necessary for the (transformative) purpose itself (Sony, ²³ Graham Archives). ²⁴ Because full-work reproduction is critical to the functioning of the search tool and the amount of text displayed in response to searchers was limited (only snippets), the court found the third factor to weigh “slightly against” a finding of fair use.

The Authors Guild argued that users could do multiple searches and access the entire work through multiple search terms and snippets, thus “replacing” for the work. The court simply dismissed this suggestion as unlikely and added that “Google does not sell its scans, and the scans do not replace the books”.

On the contrary, the court considered the fact that the search tool

enhances the sales of books to the benefit of copyright owners. An important factor in the success of an individual title is whether it is discovered --- whether potential readers learn of its existence.… Google Books in particular helps readers find their work, thus increasing their audiences. Further, Google provides convenient links to booksellers to make it easy for a reader to order a book. In this day and age of on-line shopping, there can be no doubt but that Google Books improves books sales.

Accordingly, this factor was found to weigh “strongly in favour” of a finding of fair use.

As an overall assessment, the Court concluded that “Google Books provides significant public benefits. It advances the progress of the arts and sciences, while maintaining respectful considerations for the rights of authors and other creative individuals, without adversely impacting the rights of copyright holders.” Furthermore, the Court found that Google Books advances the progress of arts and sciences by means of an invaluable research tool to efficiently identify and locate books (conducting, for the first time, full-text searches); it preserves books (specially out-of-print and old books) and it gives them new life; it facilitates access to them from remote areas and by underserved or disabled populations; and it generates new audiences and creates new sources of income for authors and publishers.

The judgment also examined two other related grounds for infringement based on Google’s actions towards the participating libraries. The District Court concluded that fair use could also exempt Google from providing the participating libraries with scanned copies of their books. According to the court, Google simply provides a means for these libraries to obtain a digital copy of a work they already own, to carry on lawful uses “consistent with copyright law” and even in other “transformative ways” such as preservation, full-text searchable indexes, access by disabled users, etc. The claim for secondary liability against Google also fails to the extent that library actions are protected by the fair use doctrine (here the court turned to the HathiTrust ²⁵ case where massive scanning done by a library was deemed to be a fair use): “If there is no liability for copyright infringement on the libraries’ part, there can be no liability on Google’s part.”

Even though the result may look similar, fair use is not an exception or limitation to exclusive rights but rather a defence against a claim of copyright infringement. As a defence, fair use is an equitable rule of reason, which can only be examined and decided on the specific facts of each infringing case. No single factor will determine whether the use is fair or not, and all must be weighed together in light of the particular circumstances of each case. The fair use doctrine was codified for the first time in section 107 of the 1976 USCA adopting the set of standards historically developed by courts to balance equities in copyright infringement claims. Its statutory formulation was not intended to limit or otherwise alter the scope of the fair use doctrine, which remains a rule of equity. For this same reason, it continues to be a critical tool to accommodate copyright in the evolving technological markets, especially where copyright laws fail to envision exempted uses.

What is interesting in this ruling is, to may view, not so much the analysis of the fair use factors, which is quite orthodox, but rather the sound recognition of the public interest of the Google Book project and the reminder that fair use – as a defence to a claim of copyright infringement – is aimed at ensuring “copyright’s very purpose: to promote the progress of science and useful arts”. In other words: Fair use as a guarantor of copyright!

Any ruling in favour of fair use only allows for the specific circumstances of the case, as considered at the time of the judgment. For instance, the Court concludes that “Google does not sell its scans, and the scans do not replace the books”, and it is on this basis that the finding of fair use was entered. If ever Google does sell its scans (and it is clear that Google would like to do that someday, as already agreed with the publishers) or do anything beyond/different from the specific circumstances now considered, it may need a license from the rightholders or a new shelter under the fair use doctrine. In fact, many of the claims raised by the plaintiffs were based more on eventual actions by Google or potential damages (“what if?”) rather than on current circumstances.

And, last but not least, uses deemed fair are not compensated. The uncompensated nature of fair use severely trims down the scope and flexibility of this defence, but it helps preserve the very nature of the exclusive rights granted to authors. Balance must be struck somewhere, and the fair use doctrine has proven to be a flexible and useful tool to achieve it.

The finding of fair use in this case should not come as a surprise. First, because it follows from a very orthodox exam of the four statutory factors and the previous case law – Google was carefully advised on the contours of the fair use defence to design its powerful tool. Second, although some see a shift in position between the two rulings, the truth is that the ASA would have granted Google far more rights (and over more works) than the fair use ruling does. ²⁶ Besides, Judge Chin already acknowledged in the ASA ruling that “the digitization of books and the creation of a universal digital library would benefit many”.

Fair use or not, it is indeed hard to deny the public interest of Google Books, an amazing tool for the advancement of the arts and science, as the judge stated. Can anyone claim that this project should not be done when technologies make it possible and easy? Can anyone claim that society should not benefit from access to any published work, spreading knowledge and information at the click of the mouse? Can we afford copyright to become an obstacle for the spread of culture? This is precisely what the fair use doctrine is envisioned to do: to prevent copyright from becoming an obstacle for cultural development. Nothing more, nothing less!

It is very unlikely that a similar result could be achieved under any EU law. Fair use does not exist in Europe on a general basis, and it is unlikely that any of the existing limitations or exceptions listed in national copyright laws could exempt Google’s actions in this project.

The quotation limitation in Article 5(3)(d) ISD ²⁷ might allow for the showing of snippets resulting from searches. It covers both the rights of reproduction and communication to the public (including the making available online); it is open-ended as to beneficiaries, purposes (the wording “such as” means that “criticism or review” are listed as mere examples) and as to the extent and nature of the quoted works. Similar limitations exist in all national laws, albeit sometimes with a more restricted scope which could hamper (if not stall) the exemption of snippets at all. ²⁸ However, quotation limitations would hardly ever allow for the whole scanning, indexing and storing of the book, which would fall under the wide scope of the exclusive right of reproduction in Article 2 ISD.

Of course, Article 5(1) ISD ²⁹ exempts the temporary acts of reproduction which are transient or incidental and an integral and essential part of a technological process, whose sole purpose is to enable (a) a transmission in a network between third parties by an intermediary, or (b) a lawful use of a work, and which have no independent economic significance. Google might argue that the snippets and extracts shown on the Google Books search tool have been automatically generated by the search engine, have no independent economic significance (rather than perhaps fostering a subsequent purchase of the book identified which would ultimately benefit the copyright owners) and that they are a lawful use (despite not being licensed or directly allowed by a limitation or exception). Even in light of the ECJ rulings in the Infopaq case, ³⁰ the display of snippets might qualify as temporary and transient act of reproduction; but once again, the scanning, indexing and storing done by Google is far from being temporary, transient or incidental.

One may then wonder whether there is also still room for “mere use” ³¹ and/or “non-substantial” reproduction online, or whether the only acts exempted from the broad scope of reproduction in Article 2 ISD are the restrictive “temporary, transient and incidental” derogation of Article 5(1) ISD. Perhaps Article 5(1) ISD could be interpreted more widely, aligning the requirement of “no separate economic significance” with the scope of the “lawful use,” in the sense that “if a specific use of a work is lawful, technical reproductions necessary to enable such use should be deemed as not having independent economic significance”. ³² The application and interpretation of any copyright statutory provisions must necessarily allow for some flexibility to take into account considerations of equity in the specific circumstances of the case, especially at a time of technological change and with copyright laws which fail to envision all the nuances of new technological uses and markets.

Otherwise, failing any equitable interpretation of the existing statutory limitations, the European copyright tale for Google Books is just the opposite of the one reached in the US. The French case Les Editions du Seuil ³³ is a good example. The Syndicat National de l’Édition (SNE) and several individual publishers sued Google for copyright infringement on the Google Books project; Google alleged in its defence the quotation limitation in the French IP Code, but the Court refused it because the works “are made available to the public in their entirety, even in reduced form, and the randomness of the choice of excerpts displayed denies any informatory purpose as required by Article 122-5-3 CPI”. ³⁴ Being a French case, it is at least surprising that the Microfor ³⁵ ruling was not mentioned to support that Google’s actions (both the scanning and the snippets) could be exempted as quotations for informatory purposes. Of course, this ruling was issued years before a restrictive reading of the exceptions and limitations (namely, through Article 5(5)ISD) was forced into European national laws, at the time when copyright limitations were interpreted (like any other provision in the copyright statute) according to general hermeneutical rules (such as the meaning of the words, the legislators’ intent and the goal to be achieved).

Certainly, even though a limitation existed that could formally allow for the unauthorized acts done in the Google Books project, compliance with the three-step test as currently applied in the EU would likely defeat its exemption. The current three-step test in Article 5(5) ISD is nothing like the fair use doctrine. Perhaps the original three-step test in Article 9.2 Berne Convention was (and remains) an enabling tool addressed to national legislators to correctly balance and design the scope of new limitations and exceptions. ³⁶ But as it has been enshrined in Article 5(5) ISD, the three-step test appears now to be a hermeneutic tool with the sole intent to further restrict the public interest (usually safeguarded by the statutory limitations and exceptions). ³⁷ Fair use is aimed at allowing specific infringing uses; Article 5(5) ISD is aimed at reducing the scope of the statutory limitations and exceptions allowing for specific uses. One restricts the exercise of copyright to ensure the very goal of copyright (the advancement of culture); the other seems to restrict the scope of exempted uses (also at the expense of the very goal of copyright: the advancement of culture). Might the three-step test become the nemesis of copyright? Let’s hope not.

One may, then, wonder whether something like fair use needs to be imported into European laws. But perhaps the general principles of the law – such as the abuse of right and good faith – may play the same role in the search for equity, which should not be foreign to copyright law. This is precisely what the Spanish Supreme Court did in the Google/Megakini case: the lack of a statutory limitation or exception to allow for the use of works within the Google Search Engine was overcome by turning to the general principles of the law (such as good faith and prohibition of an abusive exercise of rights) and by means of a rather peculiar reading of the three-step test so as to impose on the copyright owner the property doctrine of ius usus innocui (a property must endure harmless uses done by third parties). ³⁸

Copyright laws are territorial, and the rules for solving applicable law to cross-border infringement of copyrights have done very little (if anything at all) to overcome this territoriality. Hence, the traditional choice-of-law rule in Article 5.2 BC: the law of the country for which protection is being sought (lex loci protectionis).

In Europe, the choice-of-law rules in the Rome II Regulation ³⁹ are slightly different:

In the French lawsuit Éditions du Seuil, the court refused Google’s argument that US law should apply, based on Article 5.2 BC, since the scanning, indexing and storing of the books took place in the US. Instead, the French court applied French law to the dispute because it was the one bearing the “most significant relationship” with the claim: French Internet users were accessing digitized French authors’ books. In fact, if this was indeed the case at trial (protecting only French works within French territory), Article 5.3 BC would have already given the answer: protection in the country of origin is subject to its law.

Ultimately, whether under the lex loci protectionis or under the law closest to the case, several national IP laws will end up applying to Google Books claims in different countries, and with different outcomes. The lex loci damni (law where the damage occurs) would also fail to overcome the application of several national laws, since the authors and publishers (who suffer the damage) are nationals or residents in different countries. Even splitting the case in two – upload (scanning, indexing and storing) and download (searches by users) – the applicable laws (as well as the likely outcomes) would remain territorial and multiple.

In short, current choice-of-law rules may lead to several national laws to examine the Google Books project. And yet Google has relied only on US copyright law (and the fair use) to develop it and to market it all over the world. Google Books is a good example for questioning the legitimacy of territorial IP laws in the online environment. Lacking a system of harmonized national copyright laws (even within the EU market), choice-of-law rules based on one single applicable law (lex loci originis) instead of on multiple applicable territorial laws (loci protectionis/loci damni/closest connection) are an absolute necessity.

Legal uncertainty ultimately only benefits larger agents (such as Google) who can afford the economic costs of the copyright infringement claims resulting from developing new markets and permits de facto the extraterritorial reach of a few national laws (at the expense of other applicable laws). Technological changes offer an opportunity to improve – and fine tune – European copyright laws and make sure that they remain a powerful tool for the advancement of culture. Google Books is one of these opportunities.