Select Page
Eleventh Circuit’s Fair Use Decision In Georgia State Leaves Uncertainty Over Use of Excerpts in University Course Packs

Eleventh Circuit’s Fair Use Decision In Georgia State Leaves Uncertainty Over Use of Excerpts in University Course Packs

Academia is abuzz with reactions to the Eleventh Circuit’s copyright fair use decision in  Cambridge University Press v. Patton. This is, as one blogger described it, “the most important copyright and educational fair use case in recent memory.”

The decision highlights, yet again, the truth behind Professor Nimmer’s observations that fair use is said to be “the most troublesome [area] in the whole of copyright law” and has been called “so flexible as virtually to defy definition.”

At issue is Georgia State University’s right to scan and distribute to students, without payment, digital course packs comprised of excerpts from scholarly books and journals. The plaintiffs — several publishing houses — alleged 74 specific instances of infringement. In a 350 page ruling issued in May 2012 the district court court found infringement in only 5 of these cases, handing the university a significant victory and awarding it several million dollars in attorney’s fees.

The U.S. Court of Appeals for the Eleventh Circuit (which covers federal appeals in Alabama, Florida and Georgia) reversed the district court, issuing its own mammoth opinion on October 17, 2014, (129 pages including the 16 pages of concurring opinion by one of the three judges on the panel).

Needless to say, this case is of great importance to colleges and universities that want to provide their students with course packs without having to coursepackspurchase “permissions” from publishers. It is, in effect, a test case based on 74 individual claims of infringement. The district court evaluated each work individually, taking into consideration factors such as the amount of copying (a finding that the school copied less than 10% or one chapter or less favored fair use) and whether the work was available for digital licensing (licensing unavailability favored fair use). At the end of the day, the district court found that all but five of the 74 items were protected by fair use.

The Eleventh Circuit agree with many conclusions reached by the district court –

  • It agreed that the works should be evaluated individually, rather together as a “nebulous cloud of infringements.”
  • It agreed that Georgia State’s use of the works was of a nonprofit, educational nature, a conclusion that heavily favored fair use.
  • It agreed that license availability should be taken into account in evaluating the effect of the use on the market for or value of the excerpts. (Although it found that the district court did not give this factor proper weight).

However, it disagreed in some important respects:

  • The Eleventh Circuit disagreed that the “educational and informational” nature of the excerpts favored fair use, concluding that in some instances this factor was neutral, or weighed against fair use.
  • It rejected the district court’s mechanistic use of a “blanket 10 percent-or-one-chapter benchmark” as a “substantive safe harbor.” The district court “should have analyzed each instance of alleged copying individually, considering the quantity and quality of the material taken – including whether the material taken constituted the heart of the work – and whether that taking was excessive in light of the educational purpose of the use and the threat of market substitution.”
  • It disagreed with the fact that the district court gave each of the four statutory fair use factors equal weight – the court should have used a “holistic analysis which carefully balanced the four factors.”

After this detailed critique of the district court’s decision, the Eleventh Circuit sent the case back for “further proceedings consistent with this opinion.” In other words, the district court must now evaluate each of the 74 excerpts again, applying the criteria and methodology described in the Eleventh Circuit’s opinion.

coursepacks2To complicate things further, one of the three judges on the 3-judge appellate panel wrote an opinion which, while technically concurring (since it agreed to remand the case to the district court), is for all practical purposes a dissent. In the view of this judge (a district court judge sitting by designation) the case involved “extensive verbatim copying in undisputed non-transformative format, resulting in complete market substitution”. The “concurring” judge noted that Georgia State had paid permission fees when distributing the articles in paper format, and argued that the “use” of the copyrighted works did not suddenly become fair use “just because the work is distributed via a hyperlink instead of a printing press.” He concluded his opinion with the statement that Georgia State’s use of the publishers’ copyrighted works “without compensation was, in a word, unfair.”

What are some takeaways from this decision?

First and foremost, the case illustrates, yet again, how difficult it is to apply the fair use doctrine in practice. The four judges who wrote on the case represent three different views on the application of fair use to these facts. Georgia State could have lost the case outright if one of the two judges who backed the Eleventh Circuit “majority” opinion had agreed with the “concurring” (actually dissenting) judge. That is a thin margin of victory.

Second, as a “test case” the outcome of the appeal is not a clear winner for either side. Not only is the outcome on remand unclear (the district court must re-examine and reach a decision as to each of the 74 excerpts), but going forward the Eleventh Circuit’s “holistic” balancing approach makes it difficult for publishers and non-profit academic creators of course packs to predict whether a fair use defense will succeed or fail as to a specific excerpt. Certainly, non-profit educators would prefer a predictable, mechanical test (such as a 10% rule or the 1976 “Classroom Guidelines”*), rather than the uncertainty and expense of a holistic test.

*note: The Eleventh Circuit gave little weight to the Classroom Guidelines, noting that they are not controlling on the court, and that they contain hard evidentiary guidelines, an approach that is inconsistent with subsequent guidance from the Supreme Court.

Setting aside the test case aspect of this lawsuit, the Eleventh Circuit’s decision must be discouraging to the parties that have had to pay for this slow and frustrating process. Before this appeal, Georgia State alone had incurred attorney’s fees and costs exceeding $3 million. Both parties can expect to spend hundreds of thousands of dollars more when the case goes back to the district court and, absent a settlement, nothing will prevent the losing party there from filing a second appeal to the Eleventh Circuit. There is no small chance that the case will eventually be appealed to the Supreme Court, which hasn’t decided a straightforward copyright fair use case in 20 years.

Cambridge University Press v. Patton (11th Cir. Oct. 17, 2014)

Update: Subsequent decisions in this case:

Cambridge Univ. Press v. Becker (N.D. Ga. Mar. 31, 2016)

Cambridge University Press v. J.L. Albert (11th Circ. Oct. 19, 2018)

 

 

The Kerfuffle Over Copyrights in Pre-1972 Sound Recordings

The Kerfuffle Over Copyrights in Pre-1972 Sound Recordings

If you are confused by the news that a California federal court has ruled that satellite broadcaster Sirius XM is liable under California state law for streaming pre-1972 sound recordings by The Turtles,* you are not alone. The issues in this case prove, once again, Mark Twain’s complaint that “only one thing is impossible for God: To find any sense in any copyright law on the planet.”

To appreciate this bizarre situation you may need to be reminded of a few basic principles of our arcane copyright laws:

  1. The copyright in a musical work (aka a musical composition) and a “sound recording” of that work may be separately owned. John Lennon’s estate may own the copyright to the musical work Imagine, but Ray Charles, Elton John and Queen (or their record labels) own the copyrights in their sound recordings of the song. Thus, a sound recording may give rise to legal rights under two separate copyrights with two different owners.
  2. When Congress passed the 1976 Copyright Act, the new federal law preempted most, but not all, state copyright laws. One area Congress expressly left untouched was copyright rights in the public performance of  sound recordings made before February 15, 1972. The public performance right for sound recordings will not become subject to federal law until February 15, 2067 (go figure).
  3. The federal vacuum for pre-1972 sound recordings left by the 1976 Act gives rise to an argument that state copyright laws may be asserted by the owners of pre-1972 sound recordings.
  4. Lastly, in 1995 and 1998 Congress passed laws which have the effect of requiring noninteractive digital streaming services (such as Sirius XM and Pandora) to pay a compulsory license set by the Copyright Royalty Board and administered by SoundExchange, for post-1972 sound recordings. This compulsory licensing system does not cover analog transmissions (AM/FM radio). Nor, importantly, does it cover pre-1972 sound recordings.

Here is the rub: the owners of pre-1972 sound recordings claim that if the public performances of their sound recordings are not covered by the federal turtlescopyright act then they must, of necessity, be protected under state copyright laws.* Congress said nothing to the contrary when it passed the 1976 Act, and in fact expressly left pre-1972 sound recordings governed by state law. Flo & Eddie, which owns the Turtles’ pre-1972 sound recordings, has filed suits against Sirius XM and Pandora, as have several record companies. To date, the cases are centered in California and New York.

*note: This has implications in more than one area of copyright law. For example, in April 2013 I wrote a post discussing the question whether the Digital Millennium  Copyright Act’s “notice and take-down” process provides a safe harbor for publishers of pre-1972 works. (See New York State Court Blows a Hole in the DMCA Safe Harbors for Pre-1972 Sound Recordings).

The first court decision to address this issue was Flo & Eddie, Inc. v. Sirius XM , decided by a California federal judge on September 22, 2014. This case held that because sound recordings fixed before siriusxm_sued-297x195February 15, 1972 are not covered under the federal copyright act — and are explicitly left to the control of state statutory and common law — Flo & Eddie’s rights are governed by a California statute which vests “exclusive ownership” in pre-1972 sound recordings in their authors. The California copyright statute states:

The author of an original work of authorship consisting of a sound recording initially fixed prior to February 15, 1972, has an exclusive ownership therein until February 15, 2047, as against all persons except one who independently makes or duplicates another sound recording that does not directly or indirectly recapture the actual sounds fixed in such prior recording, but consists entirely of an independent fixation of other sounds, even though such sounds imitate or simulate the sounds contained in the prior sound recording.

Sirius XM argued that because it was generally assumed at the time the California law was enacted that copyright rights did not exist in pre-1972 sound recordings under state law, the California law was never intended to grant pre-1972 copyright holders this right. The court rejected this reasoning based on the express language in the statute, concluding that “exclusive ownership” encompasses all rights, the only exception being the right to make a “cover” recording (which is expressly carved out in the statute).

Setting aside an almost certain appeal and reversal, what are the implications of this decision? First, there is no reason that the logic in Flo & Eddie does not extend beyond streaming webcasters to analog broadcasters, such as radio and television. It’s not clear why these companies would not now be squarely in the litigation cross-hairs of Flo & Eddie and others (mostly record companies) who own pre-1972 sound recordings.

lpSecond, this decision (if upheld) puts Internet service providers such as Youtube at risk of copyright infringement under state laws. If Youtube is not protected by the DMCA (as one court has already held in a case involving a different music locker), Youtube cannot claim immunity under that section of the federal copyright law, even if pre-1972 sound recordings were uploaded by users of the site. It would be liable for copyright infringement under the same theory used in Flo & Eddie – by streaming these recordings it is publicly performing them.

Despite Flo & Eddie’s win in the federal district court in California, this battle is by no means over. First, there is an argument that Sirius XM can press on appeal, either to the Ninth Circuit or to the California Supreme Court (if certified to that court), that given the fact that public performance rights in sound recordings did not exist as a historical matter before 1972, the California legislature never intended to create them for pre-1972 sound recordings. However, this seems a long-shot given the language of the California statute. If there were any legislative history to support this argument it would have been presented to the federal district court, so apparently there is none.

Second, there are several ways the federal legislature could step in and fix this problem (prospectively, at least) by amending the federal copyright statute. In fact, the Register of Copyrights has recommended exactly that. For example, Congress could bring sound recordings (pre- and post-1972, webcasting and otherwise) under the system of compulsory licensing regulated by the Copyright Royalty Board. Legislation to effect this has already been proposed.

Third, broadcasters could negotiate licenses for these works, either directly or through a new or existing performance rights organization.

Lastly, some commentators have argued that the Flo & Eddie/record company litigation strategy vis-a-vis pre-1972 sound recordings is an attempt to obtain leverage to increase the royalty rates paid under the compulsory license for post-1972 sound recordings, which means that a compromise on post-1972 royalty rates could resolve the dispute.

The final option for streaming companies and their terrestrial counterparts would be to eliminate pre-1972 sound recordings from their broadcasts, which would certainly be the worst outcome for music lovers.

Flo & Eddie v. Sirius XM (C.D. Cal. Sept. 22, 2014)

A Contractual Acceptance Period is a Vendor’s Best Friend

I’ve often advised vendor-clients that one of the best ways to protect themselves is to include an acceptance clause in their agreements. This can be accomplished either through an explicit acceptance clause or a short warranty period, which can function as a de facto acceptance clause. For some reason, many customers seem to forget about the acceptance clause, giving the vendor a strong defense to a claim of breach.

This is what happened in Samia v. MRI Software, decided by Massachusetts federal district court judge Nathaniel Gorton on October 9, 2014.

Samia purchased computer software, consulting and technical support from MRI. The License and Services Agreement provided for a 30 day warranty period, during which Samia could notify MRI of any non-conformities and trigger a “repair-and-replacement” clause. This was, in effect, a 30 day acceptance period – Samia had 30 days to vet the software and notify MRI of any defects. This was the sole and exclusive remedy, and all other remedies were disclaimed.

The contract contained a separate provision addressing defects related to custom work authorizations. Here, Samia had a 30 day acceptance period during which it could “test any project elements … and notify [MRI] of all potential deficiencies relative to the applicable specification for such work.”

To make a long story short, Samia claimed to have found non-conformities and deficiencies, but it failed to given written notice to MRI during the 30 day period following delivery. On this basis, the court allowed MRI’s motion for summary judgment on most of its claims, leaving alive only a claim of negligent misrepresentation based on oral promises allegedly made to Samia. However, little is left of Samia’s case at this point.

If you are a vendor, do your best to include an acceptance clause (or a short warranty period) in your agreements. If you are a customer, challenge this, but if you are forced to accept it, be sure to rigorously test any deliveries in a timely manner, to avoid the defense that defeated Samia in this case.

Samia Companies LLC v. MRI Software LLC (D. Mass. Oct. 9, 2014)

 

 

Two Recent Decisions Show the Strengths and Limitations of the CDA

Two Recent Decisions Show the Strengths and Limitations of the CDA

Many observers have commented that if they had to identify one law that has had the greatest impact in encouraging the growth of the Internet, they would chose the Communications Decency Act  (“CDA”) (47 USC § 230). 

Under the CDA (also often referred to as “Section 230”) web sites are not liable for user submitted content. As a practical matter, in most cases this means Internet providers are not liable for defamation posted by users (many of whom are anonymous or judgment-proof).* 

*note:The DMCA, not the CDA, provides Internet providers with safe harbors for claims of copyright infringement based on user submitted content.

Two recent cases illustrate the reach and limitations of this law. In one case the CDA was held to protect the website owner from liability for defamation. In the other, the law did not protect the website from potential liability based on negligence.

Jones v. Dirty World

The CDA provides immunity from information provided by users. However, if a site itself is the “content provider” — for example, the author of defamation —  it is legally responsible for the publication. In other words, the CDA does not give Internet providers or web site owners license to engage in defamation, only immunity when their users do so.

Under the CDA the term “content provider” is defined as a person “that is responsible, in whole or in part, for the creation or development of information ….” Therefore, in many cases, the issue has been who is responsible for the “creation or development” of the defamatory content – the poster or the site owner?

This was the issue before the U.S. Court of Appeals for the Sixth Circuit in Jones v. Dirty World Entertainment Recordings LLC.

Nik Richie owns Dirty World, an online tabloid (www.thedirty.com). Users, not Mr. Richie or his company, create most of the content, which often is unflattering to its subjects. However, Dirty World encourages offensive contributions by its “dirty army,” and it selects the items that are published from user contributions. In addition, Mr. Richie often adds a sentence or two of commentary or encouragement to the user contributions.

Sarah Jones, a teacher and cheerleader for the Cincinnati Bengals was repeatedly and crudely defamed on the site. However, the defamation was contained in the posts written and contributed by users, not Richie or his company. In fact, it’s easy to see that Ritchie had been carefully coached as to what he can and cannot say on the site (as distinct from what his contributors say).

Dirty World refused to remove the defamatory posts, and Sarah Jones (who apparently was unaware of the Streisland Effect) sued Richie. Two federal court trials ensued (a mistrial and a $338,000 verdict for Jones).

Before and during the trial proceedings Richie asserted immunity under the CDA. The trial judge, however, refused to apply the law in Dirty World’s favor. The district court held that “a website owner who intentionally encourages illegal or actionable third-party postings to which he adds his own comments ratifying or adopting the posts becomes a ‘creator’ or ‘developer’ of that content and is not entitled to immunity.” Of course, there was a reasonably strong argument that Dirty World and Ritchie did exactly this – encouraged defamatory postings and added comments that ratified or adopted the posts — and hence the jury verdict in Jones’ favor.

After the second trial Richie appealed to the U.S. Court of Appeals for the Sixth Circuit, which reversed, holding that Dirty World and Richie were immune from liability under the CDA.

The first question before the Sixth Circuit was whether Dirty World “developed” the material that defamed Sarah Jones. In a leading CDA case decided by the Ninth Circuit in 2008 — Fair Housing Council of San Francisco Valley v. Roommates, LLC —  the Ninth Circuit established the following “material contribution” test: a website helps to develop unlawful content, and therefore is not entitled to immunity under the CDA, if it “contributes materially to the alleged illegality of the conduct.”

The Sixth Circuit adopted this test, and held a “material contribution” meant ” being responsible for what makes the displayed content allegedly unlawful.” Dirty World was not responsible for the unlawful content concerning Ms. Jones.

Second, consistent with many other cases applying the CDA, the court held that soliciting defamatory submissions did not cause Dirty World to lose immunity.

Lastly, the Sixth Circuit rejected the district court’s holding that by “ratifying or adopting” third-party content a web site loses CDA immunity: A website operator cannot be responsible for what makes another party’s statement actionable by commenting on that statement post hoc. To be sure, a website operator’s previous comments on prior postings could encourage subsequent invidious postings, but that loose understanding of responsibility collapses into the encouragement measure of ‘development,’ which we reject.”

The $338,000 verdict was set aside, and the district court instructed to enter judgment in favor of Richie and Dirty World.

The Sixth Circuit’s decision was no surprise. Many people in the legal community believed that the trial court judge was in error in failing to dismiss this case before trial. Nevertheless, it is a reminder of how far the CDA can go in protecting website owners from user postings, and adds to the road map lawyers can use to make sure their clients stay on the “safe” side of the line between legal and illegal conduct under this law.

Jane Doe 14 v. Internet Brands (dba Modelmayhem.com)

Things went the other way for Modelmayhem, in a case decided by the Ninth Circuit on September 17, 2014.

Like Dirty World, this case involved a sympathetic plaintiff. The plaintiff, “Jane Doe,” posted information about herself on the “Model Mayhem” site, a mayhemnetworking site for the modeling industry. Two rapists used the site to lure her to a fake audition, at which they drugged and raped her. She alleged that Internet Brands knew about the rapists, who had engaged in similar behavior before her attack, but failed to warn her and other users of the site. She filed suit, alleging negligence based on “failure to warn.”*

*note: The two men have been convicted of these crimes and sentenced to life in prison.

In this case, like Dirty World, the district court again got it wrong and was reversed on appeal. However, in this case the district court wrongly held that the site was protected by the CDA.

The Ninth Circuit disagreed, stating –

Jane Doe … does not seek to hold Internet Brands liable as a “publisher or speaker” of content … or for Internet Brands’ failure to remove content posted on the website. [The rapists] are not alleged to have posted anything themselves. … The duty to warn … would not require Internet Brands to remove any user content or otherwise affect how it publishes such content. … In sum, Jane Doe’s negligent failure to warn claim does not seek to hold Internet Brands liable as the “publisher or speaker of any information provided by another information content provider.” As a result, we conclude that the CDA does not bar this claim.

This ruling has raised the hackles on advocates of broad CDA coverage. Their “parade of horribles” resulting from this decision includes questioning how broadly the duty to warn extends, practical questions about how a web site would provide effective warnings, and concerns about various unintended (and as yet hypothetical) consequences that may result from this decision. However, based on the broad interpretation the courts have given the CDA in the last two decades, it seems unlikely that this case will have significant implications for CDA jurisprudence. Nevertheless, like Jones v. Dirty World, it is one more precedent lawyers must take into consideration in advising their clients.

Jones v. Dirty World Entertainment Recordings LLC (6th Cir. 2014)

Doe v. Internet Brands, Inc. (9th Cir. Sept. 17, 2014) 

 

Google Rolls the Dice, Files Cert Petition in Oracle Copyright Case

Google Rolls the Dice, Files Cert Petition in Oracle Copyright Case

Google has filed a certiorari petition with the Supreme Court, asking it to review and reverse the Federal Circuit’s May 9, 2014 decision holding that the declaring code of Oracle’s Java API software is not copyrightable. I have written about this case on several occasions, most recently on May 10, 2014 (CAFC Reverses Judge Alsup – Java API Declaring Code Held Copyrightable).

Google framed the “question presented” (framing the question in such a way as to catch the interest of the court is an art form in itself) as follows:

Whether copyright protection extends to all elements of an original work of computer software, including a system or method of operation, that an author could have written in more than one way.

The Supreme Court accepts review of approximately 1% of of cases appealed to it, and therefore lawyers spend a great deal of time and effort to make their cases as significant and interesting as possible. Here, Google asks the Court to take the case in order to decide an issue the Court deadlocked on in 1996 in Lotus v. Borland. Google’s brief states:

In 1995, this Court granted certiorari in Lotus Development Corp. v. Borland International, Inc., 516 U.S. 233 (1996), to resolve the question presented here. The First Circuit had held―consistent with the plain language of 17 U.S.C. § 102(b) but in conflict with other courts of appeals―that methods of operation embodied in computer programs are not entitled to copyright protection. This Court deadlocked, affirming by an equally divided court. Two decades later, this oft-acknowledged circuit split has deepened and the question presented has grown even more important as software has become a fixture of modern life.

This case directly implicates the unanswered question in Lotus because the Federal Circuit extended copyright protection to systems and methods of operation, including computer interfaces. That holding would obstruct an enormous amount of innovation in fast-moving, high-technology industries, in part because innovation depends on software developers’ ability to build on what has come before. If the Federal Circuit’s holding had been the law at the inception of the Internet age, early computer companies could have blocked vast amounts of technological development by claiming 95-year copyright monopolies over the basic building blocks of computer design and programming. By the time Google and countless other innovators even came onto the scene, others could have locked up the field for longer than most people will live.

. . . [By holding that] copyright protection … extend[s] to a system or method of operation so long as there was more than one way to write it the Federal Circuit usurped Congress’s role, deepened a circuit split that this Court previously granted certiorari to resolve, allowed Oracle to use copyright law to evade the limits on patent protection, and thereby blocked developers from building on what has come before. The court did so, moreover, in one of the most important cases of its kind, concerning the widely used Java language and Android platform. This Court’s review is needed now, before tomorrow’s innovation falls victim to the decision below.

Google is correct in arguing that there is a circuit split over the issue presented in Lotus v. Borland. In fact, no federal circuit court has held (as the First Circuit did in Lotus), that methods of operation based in computer software are uncopyrightable, leaving the First Circuit an outlier on this issue.

However, whether this issue will attract the Court’s attention is impossible to know – four justices must vote in favor of review, but that decision is made in the context of all of the competing cases presented to the Court. If the Court does accept review — and if the Supreme Court decides the issue it was unable to resolve close to 20 years ago — this will be an important and interesting copyright case.

The Supreme Court rarely accepts review of copyright cases, but the Court reviewed two copyright cases in its last term, so perhaps momentum will play a role in its decision whether to accept review in this case.

For those interested in how the Supreme Court viewed this issue in 1996 based on oral argument, the transcript of oral argument is linked here.