December 04, 2003
Compulsory Licensing and Digital Media

The Berkman Center at Harvard is hosting a day of discussion tomorrow on the Development of an Alternate Compensation System for Digital Media in a Global Environment. Among the attendees: Derek Slater, Ernie Miller, Ed Felten and Eddan Katz, as well as Harvard Law School Professor and Berkman Director William Fisher, who has "been investigating two possible alternatives to the current copyright system for addressing this crisis: a mandatory, state-run compulsory licensing system; and a voluntary entertainment co-operative."

Sounds really interesting. Speaking of CLs, there has been so much discussion over the past year on this, and the arguments have been all over the the map, the blogosphere, in traditional media and on various mail lists, and strongly debated, that at times it's been daunting for me to think about jumping in to comment.

But I've written a paper on CLs, located below, that is a reflection of my thoughts at the moment on this subject. My conclusion: that CLs will help us in the short term, with digital media, but long term, they may hold us back from developing new innovations for using digital media that are inconceivable today. And the burden of a new, giant copyright bureaucracy interested in perpetuating itself is also problematic, not to mention the privacy considerations, and the issues of controlling and keeping flexible such a powerful bureaucracy.

See also Ed Felten's post on these issues today.

See also this discussion on the ownership of materials on a group blog.

PS: school's winding down but the workload isn't. Sorry for the spotty posting.

Thoughts on Digital Media and Compulsory Licensing

Compulsory Licensing (CL) has been discussed for some time in academic and intellectual property circles for the purposes of finding a method to compensate artists and creators, as well as publishers and corporate works-for-hire. More recently, these discussions have centered around digital artifacts under copyright protection in the industries of music, movies, and games because these appear to be the most controversial. With the Recording Industry Association of America's aggressive actions toward those who share digital music files with other users, and the ensuing lawsuits, various systems for CL have been discussed as a solution to the impasse between music users and copyright holders, and in one case attempted in practice using the new Napster at Pennsylvania State University.

However, there is obviously more sharing of digital media on the internet than just music, and different forms of digital media may require different responses because of the number of copyright owner-participants for that media, as well as the way that media is sold in the analog world. Therefore, a single CL response for all different types of digital media seems unrealistic, at least as far as styles of compensation, and permissible uses goes, and thus means that different medias must be considered based on their properties as well as the information economics they use in the analog and digital realms. I want to look at several media in the context of CLs, including print journalism media, the kinds of uses it has now and may have in the future as it is put to uses in digital media systems, and the responses from users and participants in a CL system of this kind.

Compulsory Licensing Systems

For the purposes of this review, CL will be defined as a system that would in some way charge all users, possibly through their Internet Service Providers, through a flat fee system. Then, the fees could be distributed to copyright holders of works being shared on the internet. There are different ideas about what the distribution would look like and a key aspect of the public discussion has been about how to understand what works are being shared and how to value them in the distribution. For example, should a system track each and every user's activities, so that users exact consumption is known to their ISP, which might lead to the disclosure of a user's intellectual consumption? Privacy advocates reject this system as invasive. Should a system pay different rates depending on different

People like Julie Cohen have written that there is a "right to read anonymously", where she says that "reading is so intimately connected with speech and freedom of thought that the First Amendment should be understood to guarantee such a right." While she is addressing the right to read, I believe this same idea applies across all intellectual expression and consumption. And while it is a way of looking at freedom of thought and speech in a way most people do not conceive of, as something representing incoming expression, verses the usual way First Amendment speech is discussed in public discourse as an outgoing expression, it is important to consider this view with any CL system.

Cohen also has written about digital rights management schemes:

While a CL system would not likely restrict what users can do with the media itself that they consume from the internet (the media would, in that scenario, have some sort of DRM technology embedded into a media file itself), a CL system could be made to track users consumption and behavior in the way Cohen describes. Users' consumption habits would then be known, and thus available by subpoena or other warrant request to others. Since this kind of aggregation and tracking was not formerly so available in the analog world, some users as well as privacy advocates worry about the implications this kind of available information will have on people's habits, intellectual development and the quality and breadth of public discourse. Because of these concerns, and my belief that a unit system would cause a chilling effect on people's exploration of ideas, I believe a CL system that tracks each unit of consumption and attaches it to a particular user is a mistake.

Another version of a CL system might put counters around the internet, to look for and count media files, using some kind of sampling algorithm that would estimate the consumption across all users, and then remit CL fees accordingly. There has been discussion of these methods where the debate has centered around the ability to "game" such a system. Some of these methods include posting misnamed files for users to download, tricking the user into taking a file that is not the one desired, but causing a file to be counted nonetheless. Or files could be embedded with a hash mark, but these marks could also be forged or altered to either register no count, or some count other than the actual file it represents. Users might set up their systems to repeatedly download a particular file while they were asleep, so as to force multiple counts of files in the CL system. And of course in the music scenario, record companies and artists would have an incentive to coerce users to download particular works, or to download them themselves, in order to increase revenues. While this seems less likely with print media, it might be possible for particular groups to give incentives to their users to repeatedly download particular articles or articles generally from a particular publication.

Digital print journalism media, because there is much more of it, and because it has such a different value and meaning over time due to its value which is primarily short term with many outlets covering the same facts that are in the public domain, as compared to other digital media such as music or movies, should be considered differently. It seems unlikely that a single article, essay or even book would be copied and distributed widely the way that popular music or movie files are. Print journalism is more likely to be temporal in nature, where an article is useful to most people for a day or two, maybe a few weeks depending on the writer, subject and publisher, and then it enters an archive status, where it is accessed less often as the information in it goes out of date and new information and new articles replace it. Books, especially fiction, have more similarity to digital entertainment files such as music and film, but so far, few readers have been willing to read an entire book on screen, so readers are left to print out a digital copy. Often, printing a whole book is more expensive than purchasing it, where it comes bound, with cover art. Book piracy has not been nearly the problem that other audio/visual media has because of this. Because of these limitations with print media, and because of the volume of new digital print works available on the internet, it would seem that the incentives for gaming a CL system as described above but for this type of media would be harder and less likely.

Another scenario for gaming a CL system might be where users choose where their fee was allocated, resulting in some users allocating all of their fee to their favorite charity, say the National Rifle Association marching band, which may in fact make content that is copyrighted and downloaded, but not represent a single user's entire consumption over one period. In that kind of scenario, who is to say what content really is, for that matter, what art is, and whether one kind copyrighted content deserves to be included in a system over another. Which leads to the situation where for political or other reasons, particular content creators might be targeted for either lots of downloads, or have their content blocked. Also, in a user's choice scenario, users could simply record themselves, and remit the money back to their own account, so that they were in effect paying nothing at all, whether or not they consumed copyrighted media through the CL system.

Beyond any gaming concerns, fair use enjoyed in the analog world might be blocked in a digital media, CL system, and therefore lead to a situation where users lost the enjoyment of those fair uses in the digital world. An example might include the legitimate purchase of a media that a user might want to share with friends or family (not all 60 million friends on KaZaa, but actually friends in real life). Sharing digitally would mean all shared items would be counted in the CL system, though it's possible that some items would not be counted under an analog system. Users in the analog world might clip an article, photocopy it, and distribute it to a couple of people, but the difficulty in copying and sending it around places a natural limit on this fair use. But the ease of cutting and pasting an article into an email and sending it to a list of possibly thousands makes sharing a different and difficult issue in terms of policy. While this might seem like a relatively minor consideration, public understanding of fair use for legitimate sharing is important. Unless a shift in copyright for digital media occurred, say, encompassing a new understanding where sharing were accepted and encouraged without penalty or fee, which seems unlikely, it would be unfortunate to lose this fair use right, because the CL system simply began counting all sharing as an event causing a fee.

Some users have also expressed concern, in the one CL system currently in place, that they are not users of the content funds are being collected and distributed for, and therefore do not want to participate. Even in the example at Penn State, where fees for the Napster provided music were taken from the information technology fee assessed for students, and the overall fees were not increased though the new service was added to the system, students were aware that the funds were coming from somewhere. Eventually even if no increase were incurred due to the new Napster, they would suffer a loss elsewhere. Therefore, non-music users were offended at having to pay for something they did not want. This kind of reaction would likely be more widespread if a CL system were adopted across the internet than with the Penn State student population which enjoys a relatively high level of music downloading participation at around 55%. According to the Pew Research Center, 29% of internet users in the US download music, as of July 2003 and that number is unchanged since their previous survey in April, 2001. Pew has also surveyed the consumption of online news and determined that, for example, sites like the NY Times online has 13 million unique customers a month, and users going online for news, as of June, 2003, is around 26%. So CL solutions covering all internet users would have to contend with approximately one quarter of the user's participation verses three quarters non-participation. While these numbers might change for music under a CL situation, considering that news and music is free now online, legally or illegally, it seems that these numbers would not change that much.

All of these systems, gamed or not, would require someone to maintain a registry of copyrighted works (so that works in the public domain were not unnecessarily remitted fees) and a distribution system for allocating the collected fees. This implies some sort of government control or special entity control to make distribution decisions, police the system, and actually remit the funds. Any new system like this would require some bureaucracy to manage it, and would require some overhead to run. In effect, it would place an intermediary between the user and the copyright holder or creator, but that intermediary would be different than a commercial outfit which is subject to competition and natural pressures inherent in the commercial system to keep costs down. A single CL distribution bureaucracy, collecting all fees and redistributing them, would have tremendous power over users, the market and copyright holders, and would have huge incentive to maintain its own power, if usage changed or if a shift in the copyright regime were being discussed. Developing a new bureaucracy with such power and influence, not to mention stake in a CL system seems like a bad idea, as institutions and the individuals associated with them are loath to give up or shift their power and associated compensation.

For all of these reasons, from privacy to gaming to bureaucracy considerations, I believe that compulsory licensing would not cause more problems than it would solve, and there are other less harmful methods to address the issue of compensating creators and publishers for their work.

Digital Media

The future of print digital media will include synthesization and contextualization tools, as well as processes for granularizing and chunking content, as Newsblaster does now. This system uses natural language processing to automate a synthesization process using found news stories on the internet. Some of the effects these systems create will resemble the works that sample music and video. The future of other media, for different use and enjoyment reasons, is also going to follow this path, as computational systems for understanding and manipulating photos and video without keywords become available. These new news systems and processes will lead to a need for systems that can understand the origin of the content, both for journalistic integrity purposes as well as for attributing to and compensating artists, writers and editors, and publishers, and be able to compensate accordingly for the amount of media used, type and origin. CL systems might then be useful in accomplishing this task, though I am skeptical for the reasons addressed earlier. But there is a need to compensate creators, and I believe systems are coming that will aggregate and change content, working with users to allow the results to flow freely across the internet (verses the current attempts working against user's peer-to-peer activities). So CL solutions will likely continue to be proposed, and as the digital media develops, CL might be very useful in allowing these new systems and uses to blossom. However, longer term, CLs might then hold back technological developments, new conceptions of copyright and intellectual property that might shift what IP means, what digital media is and means, and what can be done with content. Instituting a CL system might then hurt the future of digital media, in ways we can't map now, because these systems are undeveloped and need flexibility and room to play. CLs, in a way, might cement our current copyright regime, instead of allowing it to change as digital media use develops and changes.

Rather than adopt a system that institutionalizes a large bureaucracy with heavy handed systems for tracking user's intellectual and entertainment consumption habits, as well as one that would work actively to maintain its own existence and funding, I would instead propose a policy that rewards the research and development of a P2P system that protects user's privacy, might be adopted voluntarily by creators, and compensate them in a more versatile and simple way. Research grants for these kinds of solutions could be made available to universities and other research institutions. In addition, I would also like to see a rethinking of the copyright holder's control over their media, once that media is released to the public, so that a system similar to that used in music and lyric publishing is used. Because of the ubiquity and variety of covers, the public understands the differences between artists' covers and an original work. This style of sampling, properly labeled, could be clearly understood as not part of the original work and therefore, not related to or created by the originator's interpretation per say. However, the ability of users to distinguish is not taken into account for other kinds of expressions. A change in the policy landscape that allowed the reuse of materials under copyright would dramatically change the digital media world, though some way to compensate and attribute to the creators must be adopted. Therefore, research into a system that can understand these bits of digital data, without compromising user's privacy, in order to share say, some sort of advertising or other kind of revenue, would be a market solution that might get around the various criticisms of CL systems. And a new technology or system fairly compensating them might then permit a copyright policy change for the reuse of digital media.

Posted by Mary Hodder at December 04, 2003 05:18 PM | TrackBack

i agree with your conclusion. have the implications of a system that taxes blank media sold to consumers coupled with databases of shared content by multiple organizations like ASCAP and BMI on behalf of their members to determine how the 'tax' revenue is disbursed been considered?

it would be subject to the same concerns about gaming the system, but perhaps a statistically accurate sample could be constructed of users who voluntarily give up their privacy to a 'nielsen ratings of online file sharing' type organization in exchange for being compensated in some other way. the ratings company could then provide the aggregate data to the professional organizations. and there's probably room for multiple ratings providers with different methodologies.

Posted by: dreww on December 7, 2003 02:50 AM
Post a comment

Email Address:



Remember info?