THE FIRST AMENDMENT ISSUES RAISED BY THE TROUBLING PROSECUTION OF E-BOOK HACKER DMITRY SKLYAROV
By JULIE HILDEN
Friday, Aug. 10, 2001
Increasingly, cyberlaw cases are showing that our traditional First Amendment doctrines do not work well on the Net. The recent arrest of Dmitry Sklyarov released this week on $50,000 bail for the sale of e-book-hacking software is only the most recent example of this trend.
Is the Sklyarov prosecution a classic example of copyright enforcement, or a classic First Amendment violation? What is troubling is that it is probably both.
A Classic Copyright Infringement Case?
Sklyarov was arrested because his company, ElcomSoft, sold software that allows users to disable several restrictions on the Adobe E-Book Reader most significantly, the restriction that no backup copies of an e-book can be made. Deeming the software a "device" primarily designed to circumvent copyright protection technologies, prosecutors charged Sklyarov with a violation of the Digital Millenium Copyright Act (DMCA).
As Inside.com's Roger Parloff has pointed out, due to the realities of the Internet, outlawing ElcomSoft's product may be the only way to stop large-scale infringement of e-book copyrights. Since the product converts e-books "into ordinary, unsecured PDF files," Parloff notes, the resulting files can "be distributed by anyone to anyone throughout the globe through numerous file-sharing programs." Once that happens, "[t]he cat is out of the bag."
In Parloff's view, the case against Sklyarov is simply a classic copyright infringement case. Indeed, it is copyright infringement on steroids, for the stakes are extremely high.
To see the strength of this argument, suppose that regular, hardback books contained an invisible stamp on each page that made it impossible to make a usable photocopy of the page unless a price was paid. Then suppose that a machine had been invented whose only purpose was to erase the stamp and allow users to illegally photocopy previously-stamped pages for free.
If the government forbade the use of such a machine, and even sent its maker to jail, it is likely that no one would blink. Protests like those triggered by Sklyarov's jailing would be out of the question.
Why then, one might ask, is ElcomSoft's software any different from the stamp-erasing machine? There are two basic answers to this question, both related to the First Amendment.
First, there is a serious argument that software is not merely a product (like the stamp-erasing machine), it is also First Amendment-protected speech.
Software is a set of instructions, albeit instructions to a machine. And generally the First Amendment protects the writing and publishing of a set of instructions. That is exactly what Sklyarov and ElcomSoft did in creating their software. What the First Amendment does not protect is following those instructions, if it means breaking the law. But that is not what Sklyarov and ElcomSoft did; it is what those who purchase and use ElcomSoft's product may do or, more accurately, cause their computers to do if they make e-book copies and then sell them.
By comparison, it is well established that I am protected by the First Amendment if I publish my recipe for marijuana brownies or deliver a lecture on lockpicking (or even a lecture on how to pick a particular brand or type of lock). The rationale is, in essence, that since I haven't actually bought the marijuana or picked the lock myself, I should not be the one punished.
Of course, it is also well established that I am not protected by the First Amendment if I stand on a podium and incite a crowd to immediately break the windows of all the stores behind me and loot them. At that point, the police can intervene not just against the looters, but against me.
Because I know my instructions are about to be followed imminently, responsibility for the instruction-following can be imputed to me, as well as to the looters. This is the doctrine referred to (a bit inaccurately) as the "clear and present danger" doctrine.
Software is much closer in nature to the lockpicking lecture than the looting speech. It is a how-to guide that can be used later, at any time the user chooses, not direct incitement that riles up a crowd immediately. As a result, under well-settled First Amendment precedent, software like ElcomSoft's ought to be protected even if it is ultimately used by some to break the law. Indeed, this line of precedent may be the ultimate source of the Sklyarov protesters' instinct that it is the actor who follows instructions (and not the speaker who gives them) who is the only proper culprit where the First Amendment is concerned.
Of course, while software is very similar to how-to guides that have been previously protected, the practical consequences of according it the First Amendment protection that those guides enjoy may be disastrous from a copyright perspective, as Roger Parloff suggests.
That is in part because, while software is sufficiently similar to a how-to guide to place it squarely within the protection of the precedents, it is also similar, in some senses, to a machine. Consider, for example, the hypothetical mentioned above, of the stamp-erasing machine. Does it make any sense to make a distinction, in terms of First Amendment protection, between a "manufacturer" who creates and sells the stamp-erasing machine that others predictably use to break the law, and a "speaker" who creates and sells software that others predictably use to break the law?
Perhaps not. Yet to erase this distinction would be to work a fundamental change in First Amendment law. And merely calling software a "device" as the government is implicitly doing by charging Sklyarov under the DMCA does not erase that distinction.. Under classic First Amendment precedent, instruction lists like software are speech and are protected.
Another argument in favor of the Sklyarov protesters is based on the idea of "fair use." Certain uses of copyrighted material are exempt from the copyright law, and therefore also exempt from the Digital Millenium Copyright Act, on the ground that they constitute "fair use."
What uses, in particular, are fair? The answer is closely linked to First Amendment principles.
Critical commentary that employs part of the original including, for example, book reviews, parodies, and satires is one type of fair use. Another type of fair use is that which enhances the reader's experience of the original. Certainly, for instance, researchers' notecards that quote text verbatim make fair use of the text they quote.
Arguably within this latter type of fair use is a reader's choice to make a backup of an e-book so he can make sure not to lose it (the typical reason to back up computer files), or so she can transport the e-book from computer to computer to read it where she pleases (the modern equivalent of carrying around a paperback in your purse).
This choice is, of course, exactly what ElcomSoft's software enables when it disables the restrictions on Adobe's E-Book Reader. However, the software also, at the same time, enables copyright infringement in the form of the unauthorized copying and re-sale of a given e-book to other purchasers.
That means the software is a dual-use technology, susceptible to both infringing uses and fair uses. Other examples of dual-use technologies include the VCR and, as I argued in an earlier column, Napster's peer-to-peer file sharing.
The Digital Millenium Copyright Act, by its language, clearly reaches dual-use technologies, as long as they are primarily designed to circumvent copyright-protecting technologies. But that has a number of consequences that are, from a First Amendment standpoint, highly undesirable.
First Amendment Consequences of the DMCA's Test
First, for dual-use technologies, the DMCA's test guarantees that thorny issues that will boil down to questions of the designer's intent will arise. And these questions, once raised, will be extremely difficult to answer. Inevitably, because of the DMCA, some software with significant First Amendment-protected uses will never be written or sold, due to fear of prosecution.
Second, and similarly, the DMCA's test means that the criminality of two identical pieces of software will vary merely based on their respective creators' intent.
Suppose all MP3s included copyright-protecting encryption technology, and suppose also that Napster's peer-to-peer file sharing technology had also included a mechanism that decrypted these MP3 files so that users could share them. Given the early statements of a Napster founder showing that he knew MP3 pirating would occur, the file-sharing/decrypting technology would have been found to be a blatant DMCA violation.
Yet suppose, instead, that the very same file-sharing/decrypting technology had been invented by for example a publicist trying to break his alternative rock clients into the mainstream by distributing their MP3s free online for a limited time. Suddenly, under the DMCA's test, the software would be innocuous, for it would have been designed to enable permitted, not copyright-violating, uses.
Surely it is ridiculous that the legality of two identical sets of software code could vary merely based on thoughts in their respective programmers' heads. But that is the upshot of the DMCA's test, which inevitably puts the emphasis on the speaker's thoughts and intentions exactly where, for First Amendment purposes, it should not be.
In the end, what is particularly frightening is the fact that criminal, and not just civil, penalties can be imposed under the DMCA. Software is one context where we should think twice before allowing the law of copyright to become criminal law, too.