Digital Rights Management: Defective By Design?

Yesterday, the Free Software Foundation launched a new anti-DRM initiative—in real space quite-cleverly in the yellow suits worn to manage toxic wastes and online at defectivebydesign.org (“There is no more important cause for freedom than the call for action to stop DRM from crippling our digital future”). As Peter Brown, Executive Director of the FSF put it: “In any other industry, such limitations or invasions would be considered major flaws. A media player that restricts what you can play is like a car that you won’t let you steer. Products containing DRM are defective—only, unlike other products, these defects are deliberately created by an industry that has long stopped caring about us.”

Earlier this month, I gave a talk on DRM for our alumni weekend (slides available here and a forthcoming paper here). I started the talk with a description of what we might label the three eras of copying technology. Think of these as monk time and the era of the scriptorium; the time of Gutenberg’s asymmetry; and the era of cheap copying, dated say from 1959 with the launch of the Xerox 914, the first automatic plain paper copier.

The key point to focus on defining these eras is the symmetry or asymmetry of copying costs between the author/publisher of the work and a person in possession of a single copy of the work. These are second-copy costs. Not the cost of producing the first copy, but instead, first copy in hand, how much does it cost to create a copy of that copy?

 

In the monk era—the pre-printing press era—all copying was done by hand. These were manuscripts copied one-by-one in the scriptorium. There weren’t real economies of scale in producing copies. All copies were expensive and the author/publisher, having produced one copy of the work, was no better situated to make another copy of the work than would be any holder of the work. To be sure, the technology of copying—the ability to read and write—may not have been widely distributed, so this was a key way in which copies were controlled, but presumably only the literate were much interested in copies anyhow, and for the literate, the costs of producing the second copy were high but roughly identical to the costs of the author.

 

The printing press changed all of that and introduced asymmetry in second-copy costs. The printing press obviously lowered printing costs generally, but it also introduced asymmetry as to second-copy costs. The author/publisher was much better situated than a copy recipient to produce another copy. That cost advantage served as an important way in which the effective rights of the author/publisher to control copies were made meaningful. This is not to say that we didn’t have piracy, but it was of a different sort, such as where a printer who printed the original copies of the book ran secret print runs for a pirate. (Adrian Johns’s The Nature of the Book has a good discussion of this at pages 162-64.)

 

The third era of copying is defined by cheap, symmetric copying technology. In the talk, I start in 1959 with the Xerox 914—which was far more successful than Xerox had anticipated—jump forward to Sony, with audio cassettes and the Betamax VCR in the mid-1970s, and then end with a late-1990s picture of Sean Fanning, the author of Napster, the original peer-to-peer file sharing program. These technologies decentralize copying ability and break the control defined by Gutenberg’s asymmetry.

 

Digital rights management has to be understood first and foremost as an attempt to vindicate the core exclusive right of a copyright holder, in the words of the 1909 Copyright Act, “to print, reprint, publish, copy, and vend the copyrighted work.” Changes in copying technology have severely weakened the effectiveness of that right. The Digital Millennium Copyright Act, passed in 1998, is an effort to bolster the meaningful effectiveness of the copyright holder’s core right to control the making of copies and to do so through digital rights management technology.

 

Larry Lessig is fond of saying that he would be delighted to have the copyright law that we had in the days of Richard Nixon. That law would not include the Copyright Term Extension Act or the DMCA. My guess is that the copyright industry as a whole would take that in a heartbeat, if (but only if) it came with the actual effective control over copying that existed in the 1970s, though even then that control was breaking down at the hands of Xerox. (We’d have to ask Disney where they come out, as they were the key backers behind the CTEA.)

 

The paper rights that have always been at the core of the U.S. copyright law have to be judged by their actual effectiveness on the ground. We shouldn’t be interested in theoretical rights that don’t actually implement in the real world. DRM and the DMCA are first and foremost about creating a framework to make meaningful the core exclusive right of the copyright holder to make copies.

 

We could just give up copyright entirely, and then we wouldn’t really need the DMCA either. My guess is that some on the copyleft say “yes, exactly the point.” Not my position obviously, but I understand the plausible coherence of the position. I am reading Yochai Benkler’s new book “The Wealth of Networks.” I will be interested to see how close he comes to that position; he kills off copyright for newspapers in just one page (p.40). You can also be rather old-fashioned law-and-econ about it as Mark Lemley and Tony Reese suggest: target actual copiers and boost penalties and probability of detection until copying is no longer worthwhile. Of course, that is a slow, expensive and in many ways unattractive strategy (highly disparate results and lousy publicity).

 

We can also acknowledge the difficulties of implementing a sensible system of DRM and DRM laws (the DMCA) and might conclude that the problems just are insuperable, especially if you are a darknet sort who thinks that DRM will likely be ineffective anyhow. So DRM could be used anti-competitively as a technological refusal to deal or a form of tying. In some circumstances, antitrust law targets those behaviors, and the fact that antitrust law is being violated through technology rather than paper contracts shouldn’t change how antitrust law applies. And of course we don’t bar paper contracts just because they can be misused.

 

The DMCA puts a lot of pressure on encryption researchers such as Ed Felten at Princeton and we need to figure out a sensible approach there. But again we have other regimes—that for classified information—which have to struggle with these issues, too. Difficulties in implementation aren’t necessarily dispositive. DRM also allows per-use charging and more price discrimination. Many people are opposed to that, but that isn’t an easy issue. And finally, as an alum asked me at the talk, how will DRM relate to fair use (worth a separate post).

 

But I don’t think that this is the heart of the Defective By Design critique. Instead, the point is really about whether DRM is, from the getgo, defective by design because it limits what consumers can do with the products, whether DRM is defective because, to quote yesterday’s Free Software Foundation press release, “products don’t respect the user’s right to make private copies of their digital media.”

 

Therein lies the core dispute. No such general right exists under U.S. law. Indeed, copyright has generally assigned control over making copies to the copyright owner. Not always to be sure, so that is an untextured statement, but generally. Sony would have been an easy case if there was a general private right to make copies, but as Justice Blackmun made clear in his dissent, in creating the 1976 Copyright Act, such a right was considered and rejected. Sony barely made it through 5-4—and that was for a case where the core copyrighted work in issue was given away gratis—and it isn’t clear that Sony would be decided the same way today in the current media environment.

 

The most direct consumer right to make copies was created in the Audio Home Recording Act of 1992 in Section 1008. That is limited in a number of ways—it only covers qualifying musical recordings—but, most fundamentally, all Section 1008 does is call off lawsuits for copyright infringement. That must sound like quite a bit, but Section 1008 doesn’t say anything about whether producers can limit copying through technology.

 

Cnet news.com has a story today about a London project to fit cars with devices to slow them down automatically if the car exceeds the speed limit. This will raise all sorts of issues: what if it is an emergency and I need to rush to the hospital? Consumers undoubtedly will be troubled by the devices, but might give in if having such a device sufficiently lowered insurance rates.

 

But you see the point I am sure: we can use technology and design to help enforce laws. Doing so involves lots of choices, but the fact that limits are imposed hardly makes the products defective by design. We need to decide what rules we want—about speed limits and about control over copies—and then figure out ways to make those rules work in actual practice.

 

Comments

[RandomGuy]

Excellent post, professor. I've often wondered about a system where the "keys" to the DRM were turned over to a public repository (i.e., a library) so that once something was no longer copyrighted the content could be "freed". The carrot for turning over the keys could be something like increased penalties for those who crack before the time, cf. 17 USC 402, or an increased number of years before the copyright runs out. I guess the problem is that the bulk of copyrighted material is (relatively speaking) junk and will hardly be worth anything when the copyrights run out, and who wants to store all this junk.

I still bristle when imagine a world where I can't copy something over-the-air and keep for personal use. I'm confident copyright holders have increased incentives to create in such a world; but it's hard to weigh that against what I know would be an inconvenience (how else can I get free content onto my video ipod?). These copy-lefters may be bonkers but they get the benefit of being able to point to the tangible.

[Bob]

I think we should go back to the days before the twentieth century when the music artists made their money by actually performing live and on stage. They only used their recordings as advertisements for their performances and to generate a broader fan base. Copyright law back then didn't apply to performances and recordings on the phonograph, they only applied to written works.

[Doug Lay]

You're missing a couple of important things here.

First, there is a huge difference between the copying of today and the copying of the 1970s. The difference is the speed with which copies can be disseminated globally. Today, a single cracked or leaked copy, anywhere in the world, can be distributed around global P2P networks within hours. DRM, while it may constrain 1970's-style "schoolyard" copying to an extent, is powerless in the face of the Intrnet and P2P.

Second, while anger concerning the DMCA and DRM in general is partially due to the end-run it performs around fair use rights, there are many other concerns as well. Probably of greatest concern to the FSF is the incompatibility of most if not all DRM technologies with open-source software (or "free software" as the FSF would put it). The DMCA effectively banishes open-source software from wide swaths of the consumer entertainment space. You may consider this a regrettable but minor problem; to the FSF and others it is a major injustice indeed.

There are a number of other issues with the DMCA - you can read papers by the EFF and the Cato Institute for lots more information. Increasingly, the picture is clear...DRM cannot protect content in a highly networked world, and protecting this largely failed solution (DRM) via the DMCA incurs a great price paid by consumers and technologists alike. It is past time for this bad law to go.

[Randy Picker]

Thanks for the commments.

Bob,

Help me understand this a little more. Are you against movies? TV? DVDs? Anything that involves a recorded performance?

Doug,

What I am hearing from you is that DRM can't work. That doesn't seem to address what I understand to be the heart of defectivebydesign, namely, that it is illegitimate to attempt to control copying, because consumers have the right to make copies and copyright law shouldn't control that. Is that your view as well?

On the second point, I would find it interesting to read more. Do you have relevant links?

[Doug Lay]

My view on the legitimacy of DRM is pretty laissez-faire. If content distributors want to use DRM, fine, but it should not receive special protection within copyright law...especially given that DRM doesn't really succeed in controlling copying in today's globally networked environment.

I think you can probably use Google as well as me to find relevant links, but....

http://www.cato.org/pub_display.php?pub_id=6025
http://www.eff.org/IP/DMCA/?f=unintended_consequences.html
http://picker.typepad.com/picker_mobblog/2005/08/cohen_brilliant.html

Note that the last link is from your own blog - one would hope this isn't all new to you.

[Bob]

Doug said, "The difference is the speed with which copies can be disseminated globally."

And this somehow justifies stronger copyright laws? Just FYI, I use the public library system almost every day. I have not downloaded one song from the internet (mostly because I can't rely on the quality). Instead, I check out the CD and just rip them into mp3s. Maybe the copyright laws should be changed to make libraries illegal?

Randy, I am not agianst recorded performances. What I said was that maybe artists should go back to the old model where they make money from their live performances and only use CDs for advertisements (this business model would be better form them and for us), because once the RIAA created their cartel, they stole (I mean, bought) all the copyrights to all the music, monopolized the distribution system, and ripped off both the artists and consumers. I equate the RIAA to the diamond cartel, de Beers. And I will make sure that I never buy one song from them, neither will I ever buy one diamond.

[Tim Lee]

Professor,

This is a great post that I think nicely lays out the relevant issues. I share your dislike for the notion that fair use is a positive right of consumers that must be protected at all costs. Fair use is merely an affirmative defense to claims of copyright infringement, and I believe it can be waived by contract. So in that sense, the FSF critique misses the mark.

However, I have a few disagreements:

First, if DRM is like a contract, then it should be enforced via contract law. Among other things, that would mean that the publisher would need to inform the consumer of the terms of the contract before the sale is made, the publisher wouldn't have the power to unilaterally change the terms of the contract, and the consumer would have the right to contest alleged violations of the contract in court.

Secondly, on the "defective by design" point: the problem isn't simply that consumers are denied a specific list of traditional fair use privileges. The problem is that it creates a presumption that consumers are barred from doing anything with their copyrighted content that they haven't been specifically authorized to do. That means a great many innocuous uses of copyrighted content become presumptively illegal which would otherwise have been perfectly legal. Case in point: 10 years ago, when the first MP3 players hit the market, people could "rip" their CDs in order to listen to them on their MP3 players. Today, as the first portable video devices are released, doing the same thing with DVDs is against the law.

That means that anyone who dreams up a new type of digital media device must get the permission of Hollywood or the labels before they can use any mainstream content with the device. That creates a high barrier to entry for anyone wanting to invent a new digital media device on a shoestring budget. DRM also locks open source software out of access to DRMed content, cutting off a major source of potential innovation.

Doug already plugged my Cato paper, but let me shamelessly second his suggestion. I go into all of these issues in some detail:

http://www.cato.org/pub_display.php?pub_id=6025

[Tim Lee]

By the way, on the question of whether DRM and open source are compatible, I've got a couple of posts at the Technology Liberation Front that expand on the argument in greater detail than I do in my paper:

http://www.techliberation.com/archives/038037.php
http://www.techliberation.com/archives/037844.php

[Randy Picker]

Tim,

I read your paper in the last week—in response to Doug Lay’s comment— and was going to email comments, but let me do this here instead.

I think systems competition is much more complicated than you make it out to be. So part of your paper focuses on the Lexmark DMCA case. You seem to suggest that we should want maximal competition over each component. Does that overstate your view? I certainly don’t think that that is right generally or right in the case of printers in particular. If we are plugging papers, I discuss this in great detail in my paper Copyright and the DMCA: Market Locks and Technological Contracts (online version at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=690901 ).

The contracts point is critical. You are obviously a freedom of contract sort, clear in your paper and in this comment. But you seem to think that contracts should be enforced exclusively through the court system? Yes? No? I don’t get that at all. If scientists in gov’t labs came up with a way to enforce contracts for 10% of the cost of the court system, I would hope that the government would set up this mechanism and kick contracts cases out of the taxpayer-subsidized and expensive court system. So I don’t get your strong preference for courts.

As that suggests, I think that the government is in the business of providing enforcement mechanisms for contracts and in corresponding fashion should be willing to regulate devices that have little purpose (sole purpose?) other than for breaking contracts or breaking the law. I think that that is a tad more complicated than I have just made out: we have to grapple with to what extent we want specific enforcement of contracts or 100% implementation of the law (a subject I wrestle with a bit in my post More Driving: Do We Like Automated Tickets?) And we need to figure out how much flexibility we should create and how to do it (with the automated ticket, I am stuck with ticket unless I can go to court and get the judge to overturn it). Maybe the DMCA does exactly that in creating a serious of defenses (reverse engineering, research, etc).

On your comment here, two other points. So you raise the question of giving notice to consumers. What would suffice in your view? Look at what Sony BMG did. Was that enough? (I describe this in some detail in another paper Mistrust-Based Digital Rights Management (available here http://papers.ssrn.com/sol3/papers.cfm?abstract_id=899155 )). And, as just a matter of contract, I am not sure why consumers shouldn’t be able to agree to have terms changed within some scope. Under the Uniform Commercial Code, we don’t insist that contracts be fully specified for a contract to be formed, and subsequent changes in terms aren’t that different.

The second point to make relates to the first, namely, the absence or presence of market power will matter here. So in a competitive market we should expect terms of contracts and DRM to gravitate towards those that jointly-maximize the interests of the parties. Which might those be? You start your paper (p.2) with a description of iTunes. If iTunes was able to use DRM to separate out uses—allowing some personal copying but not p2p sharing—that might very well be jointly maximizing. Yes? No?

[Tim Lee]

Professor,

Thanks for the thoughtful comments.

On the contracts point, I think it's important to distinguish two separate notions. One is that DRM is a useful means of enforcing a contract. The other is that DRM is itself a contract. If you're position is the former, I don't really have any quarrel with that--I think the utility of DRM is often overstated, but that's for the market to decide, not me.

But the argument for the DMCA, as I understand it, is that we should consider the act of circumventing DRM to be a breach of contract, regardless of the purpose of such circumvention (with the exception of the rather narrow exemptions found in the DMCA's text).

For example, if I buy a DVD, it's a violation of the DMCA for me to "rip" the DVD to play on my iPod. It's also a violation of the DMCA for me to play the DVD on most desktop Linux computers. Last time I looked, there are no contractual terms anywhere on a DVD package informing me of these limitations. Hence, the DMCA isn't enforcing an existing contract, it's treating the existence of DRM as a de facto contract, without notifying the users of that limitation.

One of the problems this creates is an information asymmetry: the consumer doesn't think to ask about possible restrictions on her use of the product until after she purchases it and discover the incompatibility. Even then, the consumer is unlikely to be aware that the incompatibility is due to a deliberate choice on the part of the publisher (backed up by the legal power of the DMCA) rather than a mere technical glitch. So the user curses the complexity of technology and doesn't think to seek out other products that might be less limiting.

More to the point, it's simply not true that most circumvention devices "have little purpose (sole purpose?) other than for breaking contracts or breaking the law," unless you're defining DRM to be a contract. A Linux DVD player, or a conversion tool to allow people to play their iTunes songs on non-Apple MP3 players are not piracy tools by any stretch of the imagination. Neither are homebrew personal video players, which are going to be effectively illegal once DRMed broadcast formats like CableCards become ubiquitous.

You ask "If iTunes was able to use DRM to separate out uses—allowing some personal copying but not p2p sharing—that might very well be jointly maximizing. Yes? No?"

My answer is that it might be, but that in practice the effect of DRM is just the opposite: it places all sorts of restrictions on innocuous uses of their content, while doing little or nothing to impede p2p sharing. (Yes, I'm the darknet sort) Indeed, many of the uses that DRM restricts are not uses the publisher could hope to charge for. Rather, they're uses that the publisher probably would be willing to grant permission for, but that because of the nature of DRM, there's no technologically feasible way to do so. For example, it's unlikely that Hollywood wanted to prevent its customers from playing DVDs on their Linux computers. Rather, the lack of legal DVD players for mainstream Linux operating systems is a result of the fact that the proprietary nature of DRM requires locking out open source projects.

As for your paper on market locks, and the economics of tying more general, I've read it, and I'm hoping to write up something more formal on the subject in the not-too-distant future, as it's a complex issue I haven't fully thought through yet. Here's where I am at the moment: the law shouldn't make it too easy to tie products together, especially in a rapidly changing marketplace in which new innovations are often introduced by small startups. And in any event, I think copyright law is the wrong tool for the job.

[Tim Armstrong]

Thanks very much for the heads-up and pointer to this post! It's a very interesting and articulate statement, even if not one with which I'm able to agree entirely. One factual quibble: I don't think the statement "No such general right ["to make private copies"] exists under U.S. law" can be made quite so categorically as you suggest. Perhaps it can fairly be said that there is no such *affirmative* right (in a Hohfeldian sense), but neither is there a clear affirmative history of noncommercial private copying being a basis for liability (for the simple practical reason that most such noncommercial private copying occurs "under the radar" without the copyright owner's knowledge). Indeed, as Jessica Litman has pointed out, we have at least the authority of Justice Stevens for the proposition that Congress did not intend the 1976 Act to reach noncommercial private copying; he said as much in the first draft of his opinion in the Betamax case. That language did not ultimately survive to command a majority of the Court, but it suffices in my view to show that the issue isn't quite as clear-cut as you have portrayed it.

[Bill]

Great blog,
Would like to learn more...

[KBillers]

If anyone is looking for good links to subject of Copyright law and DRM I suggest these sources:
www.drm.org
en.wikipedia.org/wiki/Digital_Rights_Management
www.law.berkeley.edu/institutes/bclt/copyright/about.htm