COPYRIGHT, DIGITAL DATA, AND FAIR USE IN DIGITAL NETWORKED ENVIRONMENTS
Pamela Samuelson
Professor of Law,
University of Pittsburgh Law School
Starkly divergent views exist as to the future of copyright and of fair use in a world of digital networked environments. Among those making end-of-copyright predictions are John Perry Barlow,[1] a Grateful Dead songwriter and co-founder of the Electronic Frontier Foundation, and Nicholas Negroponte, head of the Media Lab at the Massachusetts Institute of Technology.[2] Barlow's recent article in WIRED magazine, was subtitled "Everything You Know About Intellectual Property Is Wrong." This is not a very popular concept among copyright lawyers, publishers, and many authors, and it is not a position that this article espouses. Even so, it is worth noting that copyright law is a relatively recent construct in the history of the law, one that responded to circumstances arising from a particular stage in the evolution of technology for manufacturing written works.[3] As technology for producing and disseminating informational works evolves, it seems reasonable to assume that the law will continue to evolve as well. It is at least possible that the evolution will take us beyond copyright, but that obviously remains to be seen.
The view that digital technologies pose no threat to copyright law received a hearty endorsement at a March 1993 symposium on "The Impact of Digital Technologies on Copyright Law," sponsored by the World Intellectual Property Organization at Harvard Law School.[4] To stir interest in the conference, the organizers mailed to prospective attendees a set of questions that the symposium would address as well as a set of tentative answers. The mailing asked, for example, whether there was "any need for modification of the philosophy behind, and the basic principles of, the law of copyright and neighboring rights because of the advent and impact of digital technology," and went on to say that "[a] negative answer is proposed to this question."[5] Since the organizers of this symposium did not invite anyone to speak who had ever questioned the future of copyright in the digital age, it came as no surprise that virtually all of the speakers gave firm (rather than tentative) "no" answers to the questions posed, as one might expect from a program that featured a well-known set of copyright professionals and representatives of today's strongest copyright industries.[6] Virtually all of the speakers reassured themselves and the audience that copyright would continue to be as important as it is today and would not be changed by these technologies, although it might occasionally need to be strengthened.[7]
One thing I found curious about the WIPO symposium was that a number of speakers offered glowing predictions of radically new kinds of works that would be created by digital technologies; these works would be produced and distributed in radically new ways, opening up radically new market opportunities for producers and publishers of information products; the one thing that would be constant in this sea of radical change was copyright.[8]
When I later read John Barlow's article in WIRED magazine, images of the WIPO symposium speakers came into my mind when I read this passage: "Legal efforts to keep the old boat [of copyright] floating are taking three forms: a frenzy of deck chair rearrangement, stern warnings to the passengers that if she goes down, they will all face harsh criminal penalties, and serene, glassy-eyed denial."[9] Had Barlow attended the WIPO symposium, I suspect he would have considered the talks given there as prototypical instances of glassy-eyed denial.
I will leave to my readers the task of judging whether this is so, or whether the nature of digital data and of networked environments is of a sufficiently radical character that these phenomena will, in time, transform or render obsolete the copyright concept with which we are today so comfortable.[11] Weighing against the historical argument, however, is the observation that traditional copyright law has regulated the reproduction and distribution of physical copies of works (and this was true of photographs, sound recordings, and motion pictures), whereas networked environments tend to regulate access to and use of digital data.[12] Perhaps this means that there is not as much commonality between the previous technologies that copyright has swallowed and digital networked environments which many copyright lawyers are intent on helping copyright to swallow--lest copyright be swallowed by these networks.
In digital form, all copyrighted works--pictures, sounds, texts, music, or movies--consist of strings of bits. All, therefore, fit the definition of "literary work" under U.S. law.[18] This may lead to a general flattening out of the classification scheme of copyright law and the elimination of special medium-specific rules about how copyright law will be applied to particular categories of works. One potential advantage of this may be a greater simplicity in the law of copyright and a lesser need to go to copyright lawyers for a finely honed opinion on the application of some special exception.
A potentially more destabilizing consequence of digitization may arise from the fact that in digital form, works can be more than one kind of work. Bits are just bits; they don't know and don't care whether they are a literary work, a musical work, or a pictorial work until certain instructions execute that will render them as one kind of work or another. How the bits are processed will determine what kind of work will be perceived to exist. A stream of bits that was seemingly created as an audiovisual work can, for example, be processed to produce a musical work. The Canadian performance artist David Rockeby posts video cameras so that they will record the movements of persons walking into the space where the camera views converge. Rockeby's "very nervous system" produces musical sounds corresponding to the movements.[19] Neither Rockeby nor the persons who compose music by moving in the system perceives that an audiovisual work has been created; what they perceive is music. Scientific visualization programs allow researchers to assign visual symbols to different kinds of numerical data, thereby converting what would generally be viewed as a "literary work" into a "pictorial work."[20]
Bits can also be processed so that viewers perceive them to be more than one kind of work at the same time. A few years ago, I saw a demonstration of a tool to aid in the interpretation of digital data. With this tool, graphical information could, for example, be processed to produce sounds.[21] The demonstration began with viewing a chart that depicted the distribution of men (represented by one graphic symbol) and women (represented by a different symbol) in scientific fields by rank and salary. With the aid of this tool, one could assign different sounds to different graphical symbols (e.g., a deep base sound for the male symbol, a high piccolo sound for the female symbol). After programming the tool in this way, one could then run the cursor over different parts of the chart, and could thereby "hear," as well as see, the higher concentrations of men in the senior rank and salary positions. Suddenly, the chart became a kind of musical work.
This example is interesting in itself, but it also raises interesting copyright questions. If the author of the chart has not authorized use of this tool on the chart, could he or she sue the user of the tool for copyright infringement? After all, the user has arguably created a derivative work of the chart? Or could he or she sue the developer of the tool for contributory copyright infringement? I believe that the doctrine of fair use can be extended to deal with this kind of situation.[22] If the purpose of the use is research-oriented (to better understand the chart), the nature of the copyrighted work is factual (as was true of the chart), the amount of transformation is modest and transitory, and there is no significant potential for harm to the market for the copyrighted work, it is reasonable to assume that use of the tool on the chart would be fair and noninfringing. The developer of such tools are unlikely to be held liable for contributory infringement because the tools are capable of so many noninfringing uses that it would be difficult to show any intent to contribute to infringements.[23]
Works in digital form are inherently plastic. As the singing chart illustrates, works in digital form are easily transformed from one form to another. They are also easy to alter, even without a change in the character of the work. With the right tools, it is easy to digitally edit photographs, texts, or anything else--and to do so in a manner which cannot easily be detected. (This is one reason why photographs, video tapes, and even documents may be less useful as evidence than they once were. It is simply so easy to alter them and make what appears to be a perfect original.[24]) Because of this, it is clear that the digital medium shifts somewhat the balance of power between authors and publishers, on the one hand, and consumers of copies of copyrighted works on the others. Consumers are no longer passive recipients of a physical text which they can only read or look at. Armed with digital tools, consumers are now users capable of taking an active role in interacting with texts to enhance the usabililty of them.
There will surely be a need to restrict some plastic uses of works in digital form as copyright infringements. However, fair use and fair dealing doctrines may be increasingly useful as flexible instruments with which to balance the interests of authors in the continued fixity of their works and the interests of owners of copies of copyrighted works who may want to take advantage of digital tools to enjoy the plasticity of works in the digital medium, just as it will to deal with other situations arising in digital networked environments, such as making noncommercial copies of some copyrighted texts distributed on the Net.[25] Countries that do not have fair use or fair dealing doctrines may find it more difficult to adapt their copyright laws to dealing with questions posed by digital technologies because they lack a balancing mechanism of this sort. Since it is necessary to make copies of digital works in order to use them, a copyright law that regards all uses of digital versions of copyrighted works as infringements may be too rigid to be enforceable or to command respect.
Yet another automatic generation technology that is already important is that which permits the automatic creation of databases. It is possible to program a computer system to detect certain phenomena at certain intervals and to record this information in digital form. Databases of this sort, consisting of geological and other data, have become increasingly important for commercial as well as for scientific purposes.
Programs that can automatically generate other works pose difficulties for copyright law principally because this law has always assumed that protectable works will have human "authors" who were directly responsible for every word, line or note of which the works consist. Although the automation responsible for their creation means that no human author can readily be designated for such works, commentators have been reluctant to say that no one should own rights in them because of the commercial value the automatically created work may have.
Some U.S. commentators have argued that copyright can be used to reach products of automatic generator programs.[28] Continental Europeans who take copyright's "originality" standard very seriously can be expected to press for a "sui generis" (of its own kind) solution to this problem. Indeed, the Europeans have already proposed a new form of intellectual property law for automated databases.[29] Even England which has no difficulty offering copyright protection to works lacking in creative authorship on account of the "sweat of the brow" that must generally be invested to bring uncreative works into being (such as a data compilation) has a sui generis provision on computer generated works, giving the user responsible for generation of the work 25 years of rights to copy and commercially exploit it.[30]
Whether these works are eventually brought within the body of copyright law or protected by a sui generis regime, there will very likely need to be some sort of fair use or fair dealing limitation on the scope of the intellectual property rights in generated works, so that users who are not rightsholders can make reasonable uses of them. Indeed, the proposed EC database directive would give users of unoriginal databases a right to make fair extractions from the database contents.[31] This provision is notable, in part, because most European countries do not have general fair use provisions in their copyright laws.
One promising body of research aimed at addressing this problem focuses on the development of technological means for protecting copyrighted works in digital form.[33] Some researchers are studying systems for encryption of the contents of digital works, with decryption available only after payment has been received.[34] Some are investigating the embedding of digital signatures in seemingly unimportant portions of digital pictures or movies, so that regardless of the digital manipulation that might be done to parts of these works, it would still be possible to detect that the subsequent work derived from the first work.[35] Perhaps digital works will eventually be able to report back to their rightsholders if the user is abusing access rights in these works.
Another promising area of research focuses on the potential for increased use of "header contracts" in digital networked environments.[36] Most of the commercial electronic information providers today already rely heavily on contracts and control over access to the digital information they purvey as a way to guard against market-destructive copying of their assets. In digital networked environments, a user may send out a request for access to a certain kind of information. In the future, intelligent agents operating either for information providers or for the user could search various repositories on the Net to determine whether information corresponding to this person's needs is available, and if so, on what terms. The intelligent agent could then forward to the user the various "headers" sent to it by information providers. If the user replied to a particular header as a means of ordering the information, the user would thereby have bound him- or herself to the terms of permitted use designated in the header. Software developers like the idea of such contracting with customers, for it would seem to overcome the enforceability problems that have beset their efforts to limit uses by means of "shrink wrap licenses" that so often come with packaged software sold in mass market outlets.
If copyright comes to an end in the digital networked environment, it will likely be as a consequence of the use of technological means for protecting digital works in networked environments and of header contracts. If works are protected against unauthorized copying by means of technology and contract law, there may be nothing for copyright to do, except perhaps to serve as a kind of deus ex machina justifying the use of technological and contractual means for protecting works in digital form. Even if copyright continues to serve at least a ceremonial role in the protection of texts and other works in networked environments, header contracts and technological protections of digital works in networked environments may spell the end of fair use. After all, the header will have informed the user about what uses can be made of the work, and the user will have contracted to use no more than this. Encryption of the work may also make it impossible to borrow any part of someone else's work without paying for it.
It would be ironic if the principal role that copyright came to serve in digital networked environments was to protect encrypted works against decryption (that is, ensuring that copyrighted works cannot be read rather than, as has been its traditional function, to promote the dissemination of knowledge). Any effort to decrypt a digital work would necessarily require intermediate copying of the encrypted version. Such copying would arguably be beyond the decryption privilege established in the U.S. in the Sega v. Accolade case[37] and would certainly be beyond the decompilation privilege established in the European directive on computer software protection.[38]
To some degree, the impact will depend on public policy decisions that will be made in the coming years. The outcome of these decisions will, in turn, depend on the willingness of those who favor broad user rights to formulate proposals that will balance their interests and the interests of rightsholders. They must also engage in political action to gain broad support for their proposals. If the history of copyright legislation teaches any lesson, it is that those whose commercial interests are deeply affected by public policy initiatives about intellectual property rights can be expected to lobby hard for any initiatives that would extend their rights and against any initiatives that would limit their rights.
[2] Nicholas Negroponte, The Bit Police: Will the FCC Regulate Licenses to Radiate Bits?, 1.2 WIRED 112 (May/June 1993). Retour.
[3] The first copyright statute was the English Statute of Anne of 1710. Printing presses that had made it possible to manufacture multiple copies of books in far less time and with far less expense than had previously been possible were the technology that gave rise to the perceived need for a copyright law. See generally L. Ray Patterson, Copyright in Historical Perspective (1968). Retour.
[4] See World Intellectual Property Organization, Proceedings of the WIPO Worldwide Symposium on The Impact of Digital Technology on Copyright and Neighboring Rights (1993) (cited as "WIPO Proceedings"). Retour.
[5] World Intellectual Property Organization, General Information and Provisional Program, at 5 (Oct. 8, 1992). Retour.
[6] Among the speakers at the conference were Jason Berman, President of the Recording Industry Association of America, Robert Hadl, Vice-President and General Counsel of MCA, Inc., and Gloria Messinger, Managing Director of ASCAP. Well-known copyright professionals on the program were Jon Baumgarten and Morton David Goldberg. Retour.
[7] See, e.g., Nicholas Garnett, The Music Industry, Electronic Delivery, and Copyright, WIPO Proceedings, supra note 4, at 101-117. Retour.
[8] See, e.g., David Baron, "Digital Technology and the Implications for Intellectual Property, WIPO Proceedings, supra note 4, at 29. Retour.
[9] Barlow, supra note 1, at 85. Retour.
[10] See, e.g., Morton David Goldberg, Copyright and Technology: The Analog, the Digital, and the Analogy, WIPO Proceedings, supra note 4, at 37-54. Retour.
[11] I have previously suggested that the term "copyright" may continue to be used, even though the content of that law may change so substantially as to be unrecognizable to those who know it in its present incarnation. See Pamela Samuelson, Digital Media and the Changing Face of Intellectual Property Law, 17 Rutgers Computer & Technology Law Journal 323 (1990). Retour.
[12] See, e.g., Pamela Samuelson & Robert J. Glushko, Intellectual Property Rights for Digital Library and Hypertext Publishing Systems, 6 Harvard Journal of Law & Technology 237 (1993). Retour.
[13] See, e.g., Paul Geller, The Universal Electronic Archive: Issues in International Copyright, WIPO Proceedings, supra note 4, at 213-215. Retour.
[14] 17 U.S.C. [[section]]106(4), (5). Retour.
[15] 17 U.S.C. [[section]]108. Retour.
[16] 17 U.S.C. [[section]]114. Retour.
[17] See, e.g, Sheldon v. MGM Pictures Corp., 81 F.2d 49 (2d Cir. 1989) (novel infringed by taking of plot elements for motion picture). Retour.
[18] 17 U.S.C. [[section]]101 (definition of "literary work"). Retour.
[19] Rockeby demonstrated this system at the annual conference on computer-human interaction sponsored by the Association for Computer Machinery (ACM) in 1990. Retour.
[20] Scientific visualization programs are routinely demonstrated at the ACM's annual conference on computer graphics. Retour.
[21] For a paper discussing this tool, see Stuart Smith, R. Daniel Bergeron & Georges G. Grinstein, Stereophonic and Surface Sound Generation For Exploratory Data Analysis, Proceedings of ACM Conference on Computer-Human Interaction, at 125 (1990). Retour.
[22] This set of issues is discussed at length in Pamela Samuelson, Fair Use for Computer Programs and Other Copyrightable Works in Digital Form: The Implications of Sony, Galoob, and Sega, 1 Journal of Intellectual Property Law 49 (1993). Retour.
[24] See Don E. Tomlinson, Computer Manipulation and Creation of Images and Sounds: Assessing the Impact (Annenberg Washington Program/Comm. Policy Studies, Northwestern University, Evanston IL) (1993). Retour.
[25] See Pamela Samuelson, Copyright's Fair Use and Digital Data, 37 Communications of the ACM 21 (Jan. 1994). Retour.
[26] This set of issues has been the subject of much commentary in the law review literature, among which is Pamela Samuelson, Allocating Ownership Rights in Computer-Generated Works, 47 University of Pittsburgh Law Review 1185 (1986). Retour.
[27] See, e.g., Makoto Nagao, The Machine Translation System and Its Possible Intellectual Property Implications, Proceedings of SOFTIC's Third International Symposium on Legal Protection of Computer Software 277 (1991) (cited as "SOFTIC Proceedings"). Retour.
[28] See, e.g., Arthur R. Miller, Copyright Protection for Computer Programs, Databases, and Computer Generated Works: Is Anything New Since CONTU?, 106 Harvard Law Review 977 (1993). Retour.
[29] See Proposal for a Council Directive on the Legal Protection of Databases (1993). Retour.
[30] See Christopher Millard, Advanced Computer Systems and Their Output, SOFTIC Proceedings, supra note 27, at 439-444. Retour.
[31] See Article 8, Proposed Directive, supra note 29. Retour.
[32] See, e.g., Samuelson, Digital Media, supra note 11. Retour.
[33] In April 1993 the Coalition for Networked Information, the Interactive Multimedia Association, the John F. Kennedy School of Government of Harvard, and the Massachusetts Institute of Technology jointly sponsored a workshop on research initiatives of this sort. See Proceedings, Technological Strategies for Protecting Intellectual Property in the Networked Multimedia Environment (Jan. 1994) (cited as IMA Proceedings). Retour.
[34] See, e.g., Gary N. Griswold, IMA Proceedings, supra note 33, at 169-178 and Marvin A. Sirbu, Internet Billing Service Design and Prototype Implementation, IMA Proceedings, supra note 33, at 67-80. Retour.
[35] See, e.g., Kineo Matsui & Kiyoshi Tanaka, Video-Steganography: How to Secretly Embed a Signature in a Picture, IMA Proceedings, supra note 33, at 187-296. Retour.
[36] See, e.g., Henry J. Perritt, Jr., Permission Headers and Contract Law, IMA Proceedings, supra note 33, at 27-48. Retour.
[37] Sega Enterprises, Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992) (making intermediate copies of computer program code was fair use because the user had a legitimate need for gaining access to the information and it was available in no other manner). Retour.
[38] Article 6, Council Directive 91/250 on the Legal Protection of Computer Programs, 1991 O.J. (L. 122) 42. Retour.