This post was originally published on Prof. Katz's blog.
In April 2008, three publishers, Cambridge University Press, SAGE Publications, and Oxford University Press, filed a copyright infringement lawsuit against Georgia State University, alleging that GSU infringed their copyrights by allowing professors to upload excerpts from books onto the university’s electronic reserve system (ERes). The complaint alleged “systematic, widespread, and unauthorized copying and distribution of a vast amount of copyrighted works”, and argued that GSU “has facilitated, enabled, encouraged, and induced Georgia State professors to upload and post to these systems - and Georgia State students simultaneously to download, view, print, copy, and distribute - many, if not all, of the assigned readings for a particular course without limitation.” Unless GSU’s “infringing digital distribution practices are enjoined", the complaint asserted, "Plaintiffs, authors, and the publishing community at large will continue to face a certain, substantial, and continuing threat of loss of revenue, which will in turn threaten Plaintiffs' incentive to continue supporting and publishing the cutting-edge scholarship upon which the academic enterprise depends.”
On Friday, the US District Court issued a 350 pages decision, dismissing 94 out of the 99 claims of copyright infringement.
Of the initial 99 claims, only 75 reached the final stages of the trial. Some of the claims were dismissed on the grounds of de minimiscopying, others on the ground that plaintiffs could not demonstrate that they were the owners, while the rest were found to be fair use. Infringement was found in only 5 cases. The court found that making small free excerpts from these books available to students would further the spread of knowledge, that copying caused no harm to the sale of any of the plaintiff’s books, and that here was no reason to believe that allowing unpaid, nonprofit academic use of small excerpts in controlled circumstances would diminish creation of academic works, or would appreciably diminish the Plaintiffs' ability to publish high quality scholarly books. The court dismissed the plaintiffs’ argument that they might be forced to reduce publication of high quality scholarly works, describing it as “glib” that was unsupported by any evidence.
Despite some differences between American and Canadian copyright laws, Canadian universities should pay very close attention to this decision, not only because actual lawsuits, and judicial decisions, involving allegations of copyright infringement by educational institutions are extremely rare, but also because the reasons for the plaintiffs’ failure are highly relevant to Canada. In fact, possibly, had a similar case been brought in Canada, the result could be 99-0 for the defendants, not only 94-5.
The case should also provide food for thought for the universities that have to decide whether to capitulate to Access Copyright. It casts further doubts about the claims of the AUCC that the deal that it brokered with Access Copyright is the best possible outcome for Canadian universities.
I have had some discussions with different people in the university community, including in universities that are leaning towards signing the Model License, and it seems to me that it’s not the conviction that Model License provides any real benefits for universities that motivates them to sign it, but the fear of what would happen to them if they do not sign. This is an understandable approach. For almost two decades, most universities obtained licenses from Access Copyright, which provided some real, but mostly imaginary, comfort to university administrators that they don’t have to worry about copyright. As a result, very few universities developed any internal expertise on copyright or any internal mechanisms that would allow them to feel confident about operating without Access Copyright.
Things began to change about a year ago, when more than 30 universities decided to opt out of the interim tariff, get over their Access Copyright addiction and prepare for a new life of rehabilitation. Their plan comprised of four elements: obtaining direct licenses from publishers and other market intermediaries when necessary, relying on open-access materials when available, taking fair dealing and other users’ rights seriously, and lastly, more sophisticated risk management than the one they were used to before. In other words, many universities decided to regain ownership of their copyright lives and slowly develop confidence in their ability to do that. Even when UofT and Western decided to settle with Access Copyright, many of the opt-out universities remained resolved not to go back and confident that the AUCC would not cave in. But it did, and left many of its members confused, bewildered, under pressure to make quick decisions. Panic reaction ensued, and the option of signing the agreement—what they vowed not to do again—suddenly began looking like a comforting alternative.
From my conversations, the fear of being liable for copyright infringement plays an important role for the university administrators who lean toward signing the Model License. The fear is that despite their best efforts to secure licenses, encourage use of open access materials, and provide guidance about the scope of fair dealing and other exceptions, they may not be able to guarantee that all students and all faculty and staff will comply and some infringing copies will be made. They fear that Access Copyright might sniff around, look for evidence of infringement, and sue them. However, the GSU decision demonstrate why the fear of being sued and held liable might be overblown, and universities should conduct a much closer risk-assessment.
First, without getting into much detail, it is doubtful whether Access Copyright, who acts on behalf of copyright owners but is not the owner of copyrights, has any standing to sue for copyright infringement. Even if it has such standing, the copyright owner must normally be made a party. But the interests of Access Copyright and its members do not always align. For example, the owner may be an author, or an academic author, who has no objection for her works being used for teaching, or it may be a publisher for whom universities are important customers and suing them may or may not be perceived as good business.
Second, if contested, the plaintiff’s ownership cannot be taken for granted, and showing chain of title can be more complicated than ordinarily thought. For example, a publisher claiming to be an assignee of the copyright may not be able to produce the signed document on which the assignment depends, or it may turn out that the owner has transferred only some rights, but retained others (e.g., digital rights), etc. GSU is a case on point. 18 claims were dismissed because the plaintiffs couldn't prove ownership.
Third, the copying itself might not be infringing at all. Even the copying was not covered by a license, many if not most, of the unauthorized copies made on campus may be permissible under fair dealing or other exceptions even before Bill C-11 becomes law. Save for blatant cases of infringement, copyright owners might be reluctant to risk another CCH-like outcome in borderline cases. The GSU case confirms such prediction. It was a carefully planned case, and presumably, it would not have launched unless the plaintiffs believed that they had the best possible evidence. The fact that they lost 94 of 99 complaints is telling.
Let me now offer additional (and initial) observations on the GSU case from a Canadian perspective, which, I promise, will be interesting for my American readers as well.
Access Copyright and the Copyright Clearance Center
The plaintiffs in this case were Cambridge, Sage, and Oxford as plaintiffs, but the real entity behind the case was the Copyright Clearance Centre (CCC). When it comes to licensing educational institutions, the CCC is Access Copyright’s American little cousin. While most of Access Copyright’s income comes from the educational sector (more than 75%, according to the Friedland Report), educational licenses contributed only 12% to the CCC’s revenue. American universities, for the most part, were never convinced that the CCC could offer them anything valuable, except for being an occasional source of transactional licenses. The GSU case was brought as a test case. According to the court, CCC and the Association of American Publishers (AAP) organized the litigation, and they pay the plaintiffs’ expenses and attorneys’ fees. CCC and its counsel did the initial fact gathering concerning unlicensed copying of excerpts in the higher education community and then convinced the three publishers to join. The purpose of the lawsuit was not to recover any damages, but rather to force GSU, and convince all other universities, to obtain licenses from CCC. The availability of such licenses, what actually they would cover, was a central issue during trial. What has been revealed should be of particular interest to Canadian universities: those who already decided to sign an agreement with Access Copyright and those who contemplate doing so.
So here are some interesting facts. Like most American universities, GSU declined to obtain licenses from CCC. Occasionally it obtained some ad-hoc transactional licenses from the CCC, but the scale was rather low. In an earlier decision in this case, the court found that over a decade, GSU paid CCC less than $19,000 on license fees (i.e., less than $1,900 per year). Since GSU has approximately 32,000 students, these payments are the equivalent of less than ¢6 per student per year. Canadian universities paid Access Copyright $19 per student per year during the same period.
During trial, the CCC estimated that it would cost GSU $114,000 per year to participate in its annual licensing program (p. 29), which means that the fee per-student would be $3.56. Presumably, there is no significant difference between the repertoires of AC and CCC, because the two organizations cross-license their repertoires, but Canadian universities who sign the Model License would pay $26 per student. So when Canadian universities consider whether to sign the Model License that the AUCC claims is the “best possible outcome for Canadian universities”, they should look south of the border and think what could be the price if they weren’t so eager to be dependent on Access Copyright. This chart might help them.
Access Copyright’s Repertoire
Access Copyright’s repertoire is shrouded in mystery. This has been very convenient for Access Copyright, as well as for its licensees. Access Copyright has pretended that it can license almost every work that universities might use, and universities really wanted to believe that this was true. However, this “don’t ask, don’t tell” unwritten rule has become less stable as universities began obtaining permissions covering digital and print directly from publishers, and as many of Access Copyright’s members declined to let it authorize digital rights. As I have noted before, Access Copyright’s digital repertoire is significantly smaller than its (limited) paper repertoire. The GSU decision provides some interesting insights into the limited digital repertoire available to CCC. These insights may shed light on similar issues with respect to AC’s repertoire too.
For example, the court learned that CCC is able to license excerpts only from 60% of Cambridge's works. Oxford estimated that CCC currently (in 2011) licenses the copying of excerpts from over 90% of its titles, but there was no evidence that all of Oxford’s works in question were available for licensing through the CCC, only that some of them were. CCC offers an Academic Repertory License Service, which affords subscribers access to excerpts from about nine million titles. In 2009 digital licenses for excerpts from a million and a half (seventeen percent) of these works were available (the veracity of these general nubers was not an issue in this case). The court attributed the lesser availability of digital excerpts to the following reasons: (1) some publishers are concerned that they may not have the right to authorize distribution in digital as opposed to print format; (2) some publishers are reluctant to place digital copies of their works in the stream of commerce; and (3) sometimes publishers, for whatever reason, simply prefer limiting sales to the whole book. Cambridge did not and does not participate in this program; Oxford participated in 2009 only with its journals, not with its books. Oxford currently participates in this program. Sage participates in this program currently and did so in 2009.” (pp. 28-29) In the end, of the 46 excerpts from Cambridge's and Oxford's works, licenses to make digital copies were shown to be available for only 13 excerpts.
Canadian universities that consider singing the Model License believing it provides them wide permissions to use works in digital form should think twice, and should demand to find out what exactly will they get in return for the hundreds of thousands of dollars that they would pay AC. They should recognize that the Model License that they are encouraged to sign treats digital copying quite differently from photocopies. While AC authorizes the making of digital copies for works within its repertoire, the size and scope of this repertoire is unknown. Moreover, while AC undertakes to indemnify universities who make digital copies of works not within its repertoire, the indemnity with respect to works in “born digital format” only if the work appears on the Inclusions List (if you're surprised, read s. 13 of the Model License and the definition of the term "Repertoire Work"). To the best of my knowledge, this Inclusions List is not available.
GSU vs. CCH
The litigation strategy behind the GSU case is highly similar to that employed in other cases, notably, CCH v. LSUC. Like the CCC in the GSU case, the real entity behind the CCH case was AC, who hoped to gain a quick victory over the Law Society that would open for it new markets: law firms first, and then the broader corporate world. Like the CCC in the GSU case, AC bore the cost of the litigation. To its dismay, the CCH case ended with the worst possible outcome for AC.
An earlier case that was crucial for these two cases is the 2nd Circuit decision in American Geophysical Union v. Texaco. In that case, publishers of scientific and medical journals brought copyright infringement action against Texaco, who made unauthorized copies of copyrighted articles for use of its researchers. The 2nd Circuit rejected Texaco’s fair use defense and found in favour of the publishers. Even though, when analyzing the “effect on the market”, the court found that there was no persuasive evidence with respect to harm to the sales of subscriptions to journals, the court found that a finding of fair use would impair the ability of the publishers to license copying through the CCC, and held that the defense of fair use failed.
In the GSU case, the court accepted the Texaco approach and held that the availability of a license is relevant to the fair use analysis and would weigh in favour of the plaintiff. While the court noted the notorious circularity of the Texaco approach, it concluded that there is no ideal solution to this problem. The court’s solution seems to be treating the availability of a license as a necessary condition for defeating a fair use claim, but not a sufficient one.
In Canada, the Supreme Court avoided that circularity by rejecting the Texaco approach. Although that case is not mentioned in the decision itself, Texaco played an important role in AC’s strategy in the CCH case. AC, who was behind the litigation and granted intervener status at the Supreme Court, devoted a quarter of its intervener factum in CCH to the argument that “obtaining of licenses from collective societies constitutes a valid alternative to infringement and should be considered in the determination of the fairness of a particular dealing.” It submitted that the Court should follow Texaco and hold that “Where collective societies have created a workable market for institutional users to obtain licences for the right to reproduce works protected by copyright, courts should acknowledge that the reproduction of such works, absent a licence, will generally affect the potential market for those works, and take this factor into account in any analysis of whether a dealing is ‘”fair”.” The Court in CCH listened to argument. The Court considered the argument. The Court rejected the argument. The Supreme Court held that
“The availability of a licence is not relevant to deciding whether a dealing has been fair. As discussed, fair dealing is an integral part of the scheme of copyright law in Canada. If a copyright owner were allowed to license people to use its work and then point to a person’s decision not to obtain a licence as proof that his or her dealings were not fair, this would extend the scope of the owner’s monopoly over the use of his or her work in a manner that would not be consistent with the Copyright Act’s balance between owner’s rights and user’s interests.”
Since the Supreme Court of Canada foreclosed a Texaco-style finding of infringement, the 5 successful claims in the GSU case might have failed in Canada like the rest of the 94.
Fair Use, Fair Dealing, and the Effect on the Market
I have recently blogged about the false claim that the “effect on the market” is the preeminent factor in fair use analysis in the US. The GSU case confirms this point: “It is hornbook law that there is no across the board rule for what weight should be given to each factor or how the factors should be applied.” (p. 19)
In that post I also discussed the fact that publishers have consistently failed to demonstrate any harm to their markets in virtually every reported case. The GSU case has now become just the latest chapter in a series of cases in which publishers make passionate claims about the harms of unauthorized copying for research or teaching but fail to support them with any credible evidence. As Judge Evans put it, these arguments are glib.
GSU, CCH, and the University’s Liability
The named defendants in the GSU case were university officials. In an earlier decision the court held that they could not be liable for direct infringement or under vicarious liability but left open the possibility of contributory liability. At the close of Plaintiffs' case, the court granted Defendants' motion for judgment on Count II (contributory infringement), leaving only the claim under Count I that the 2009 Copyright Policy caused copyright infringement.
The court concluded the university was liable for the five cases of infringement, because GSU’s Copyright Policy “did not limit copying in those instances to decidedly small excerpts as required by this Order. Nor did it proscribe the use of multiple chapters from the same book. Also, the fair use policy did not provide sufficient guidance in determining the “actual or potential effect on the market or the value of the copyrighted work,” a task which would likely be futile for prospective determinations (in advance of litigation). The only practical way to deal with factor four in advance likely is to assume that it strongly favors the plaintiff-publisher (if licensed digital excerpts are available)." Still, the court believed that “Defendants, in adopting the 2009 policy, tried to comply with the Copyright Act. The truth is that fair use principles are notoriously difficult to apply. Nonetheless, in the final analysis Defendants' intent is not relevant to a determination whether infringements occurred.” This is potentially another point which could turn out differently in Canada.
Normally in a similar factual setting in Canada, the university’s liability would depend on whether it would “authorized” the infringement. In CCH the Court adopted an approach that is likely to make it more difficult to hold the university liable. The Supreme Court emphatically emphasized that
While authorization can be inferred from acts that are less than direct and positive, a person does not authorize infringement by authorizing the mere use of equipment that could be used to infringe copyright. Courts should presume that a person who authorizes an activity does so only so far as it is in accordance with the law. This presumption may be rebutted if it is shown that a certain relationship or degree of control existed between the alleged authorizer and the persons who committed the copyright infringement.
Even though the Court relied on the lack of master-servant or employer-employee relationship between the Law Society and the patrons to support its conclusion that the Law Society lacked the necessary control over the library’s patrons, concluding that the existence of such relationships is sufficient to rebut the presumption that the activity was authorized only so far as it is in accordance with the law is would not only be a logical fallacy but it would also ignore the policy rationale behind the Court’s holding. In CCH the court declined the follow the Australian High Court decision in Moorehouse (which held that a university was liable for authorizing copying by library patrons becuase it failed to had taken reasonable steps to prevent an infringing copy from being made) stating that Moorhouse approach was “inconsistent with previous Canadian and British approaches to this issue,” but more importantly, it “shifts the balance in copyright too far in favour of the owner’s rights and unnecessarily interferes with the proper use of copyrighted works for the good of society as a whole.”
Moreover, despite being an employer of professors, the university is an inherently non-hierarchical organization. This is not a bug, but a feature. The very limited control that university administration has over professors choices of materials used for research or teaching is, in part, necessary to enable academic freedom, which is, as the Alberta Court of Appeal just confirmed, “an important value in Canadian society” and a value that is inextricably linked to freedom of expression protected by the Charter. Academic freedom, of course, is not a free pass to professors to infringe copyright, but a university that would be liable for the infringement of its faculty—except in some extreme situations—would have to exert much greater level of control over its professors than it currently does. An interpretation of the Copyright Act that requires a high level of control runs the risk of being constitutionally invalid.
In addition, in CCH, the Court rejected the notion that the Law Society had to prove that all the copies made by the library’s patrons would constitute fair dealing. The Court held that it was not incumbent on the Law Society to adduce evidence that every patron uses the material provided for in a fair dealing manner, but that instead, it was sufficient for the Law Society to rely on its general practice to establish fair dealing. “Persons or institutions relying on the s. 29 fair dealing exception need only prove that their own dealings with copyrighted works were for the purpose of research or private study and were fair. They may do this either by showing that their own practices and policies were research-based and fair, or by showing that all individual dealings with the materials were in fact research-based and fair.” (para. 63). The policy of the Great Library which the Court upheld as fair is much more general than the GSU policy, and does not seem to have the degree of specificity that Judge Evans requires in GSU.
The Canadian Paradox
As I write this post, Bill C-11 is under third readings at the House of Commons. Barring some unexpected development, "education" will be explicitly added to the fair dealing provision, removing any doubt that the Copyright Act allows fair dealing for this purpose. Bill C-11 will also reduce the risk of excessive statutory damages for non-commercial activities. The cumulative effect of this amendment the the Supreme Court's holdings in CCH, will be an expanded ability of Canadian universities to pursue their mandate without being subject to Access Copyright's de facto Education Tax, and without being liable for copyright infringement. To Supreme Court's clear repudiation of Texaco gives Canadian universities a wider zone than that available to their American counterparts. Nevertheless, American universities are much more willing to assert and defend their rights, while many Canadian ones, short-sighted, extremely risk-averse, and ill-advised, still cling to their habit of being dependent on Access Copyright.
Still, the opposition to the AUCC-AC Model License continues to mount. The times they are a-changin'.