The Authors Guild, representing the interests of authors in the United States, has publicly alleged that Amazon's recently-released e-book reader, Kindle 2, violates an author's right to control creation of derivative works. Kindle 2 has a text-to-speech feature that enables the device to read aloud. Certainly, text-to-speech programs are nothing new to authors, as they are often used by the blind. Why then did the Authors Guild object to Kindle 2 that employs a similar technology?
The concern is that the reading feature will supplant audio book purchases, as an audio book generally costs more than a digital e-book. In the Guild's view, the feature allows e-book purchasers to have the equivalent of both a hard copy and an audio book in one digital file. The Guild's objection appears to be that the text-to-speech function, when utilized, creates a derivative work from the digital version of the written work in violation of the author's bundle of rights.
Whether a derivative work is created depends on whether or not Kindle's text-to-speech function actually creates a work fixed in a tangible medium that can be governed by copyright laws. If it does not, the Authors Guild's argument may not hold water, because the function would be equivalent to another person reading the text aloud (though not as a public performance).
In considering the issue of fixation the case law surrounding copying can be applied. In a recent case, Cartoon Network v. CSC Holdings, 536 F.3d 121 (2nd Cir. 2008), the court held that when just over one second of an audiovisual work, a "buffer," is held on a device such as a remote-storage DVR, the copy is only transitory and does not violate copyright holder's rights. It is important to note that Cartoon Network narrowly construed, but was not decided contrary to, the existing doctrine and precedent in MAI Systems Corp. v. Peak Computer, Inc. 991 F.2d 511 (9th Cir. 1993). In both cases the courts considered what constituted unlawful copying and whether there was sufficient fixation to bring a copy within the purview of copyright law. The same consideration would be needed if the Kindle controversy were litigated—is text-to-speech "copy" sufficiently fixed or is it merely transitory?
In MAI, the court considered whether loading a program into RAM makes a copy of the software that can be restricted by copyright. The Ninth Circuit decided that since such copy could be "perceived, reproduced, or otherwise communicated," the loading of software into RAM created a copy that infringed plaintiff's copyright in the software (In 1998, §117 of the Copyright Act was amended to create an exception allowing copies at issue in the case). In distinguishing this decision in Cartoon Network, Second Circuit reasoned that there were two requirements to find the fixation mandated by the statute, not just the single issue of embodiment considered by the Ninth Circuit in MAI.
To be "fixed," as defined in the statute, the work must be embodied in a way that is "sufficiently permanent or stable to permit it to be perceived, reproduced, or otherwise communicated [the embodiment requirement] for a period of more than transitory duration [the duration requirement]" (17 U.S.C §101).
Hence, Second Circuit's recent ruling in Cartoon Network is not inconsistent with the decision in MAI, rather it narrowly construes the prior ruling and distinguishes the facts at issue. The court states that it did not interpret MAI decision to stand for the proposition that "as a matter of law, loading a program into RAM always results in copying." In fact, the Second Circuit notes that the court in MAI did not consider the duration requirement because it had already been established that the copy at issue was embodied in RAM for "at least several minutes" leaving the question of duration not in dispute (536 F.3d at 128). Because in Cartoon Network the transitory nature was in dispute, the court could not be rely on the prior decision in MAI in its entirety.
The decision in Cartoon Network provides some guidance as to what is not long enough to create a new work or an unlawful copy of an existing work, but it does not set a standard for duration that would make a copy more than a transitory fixation. Such question remains open, and if Kindle's text-to-speech function were considered by a court, it would lend further interpretation to MAI and Cartoon Network decisions with regard to duration and fixation. However, if the output of the text-to-speech function is also only "transitory" like a DVR buffer, then, consistent with the rule in Cartoon Network, no new work is created by Kindle's text-to-speech function, meaning it cannot be found to violate author's copyright.
The definition of what constitutes a derivative work is also not settled. The Ninth Circuit found in Mirage Editions v. Albuquerque A.R.T. Co., that a copyright owner could essentially block any use or display of the copyrighted work (in that case, the transfer of a printed work onto tiles), the Seventh Circuit later found otherwise in Lee v. A.R.T. Co. in a highly similar set of facts. See Mirage Editions v. Albuquerque A.R.T. Co., 856 F.2d. 1341 (9th Cit. 1988), Lee v. A.R.T. Co., 125 F.3d 580 (7th Cir. 1997). While the Seventh Circuit opinion is the more widely accepted result, both precedents have been established.
If Kindle's text-to-speech function does not create a fixed digital recording it may also be relevant to consider whether using the feature to "read aloud" constitutes public performance, which is also controlled by copyright. To be a public performance, the act must to occur in- or be transmitted to a public place, which is anywhere there is a "substantial" number of people beyond one's own family and acquaintances. Hypothetically, Kindle's text-to-speech feature can be easily used for public performance within the meaning of the statute, though the intended purpose of the function is personal convenience—"In the middle of a great book or article but have to jump in the car? Simply turn on Text-to-Speech and listen on the go"—and as we learned from the Sony doctrine and its progeny, the mere fact that new technology has the ability to be used in an infringing manner is not enough to find liability for contributory infringement, provided there is a substantial non-infringing use.
Beyond the assertions of infringement and unlawful creation of a derivative work, one wonders if the Authors Guild's concerns stemmed primarily from the fear of lost audio book sales. Most audio books are read by professional voice-over artists or actors. The performance is a strong part of an audio book experience and allure, one that is not likely to be replaced by the computerized voice of Kindle 2.
In response to the Authors Guild, Amazon has agreed to allow authors to "opt out" of activating the feature on a case-by-case basis, while asserting without equivocation that it firmly believes the text-to-speech function is legal.
Vol. 53, June 2010