May 30, 2001
Hon. Jose Cabranes
Hon. Jon Newman
Hon. Alvin Thompson
United States Court of Appeals
for the Second Circuit
40 Foley Square
New York, NY 10007
Pursuant to the Court's invitation, 2600 Magazine responds to the questions asked by the Court in its Order of May 8, 2001 and to two issues that arose during the oral argument of this case on May 1, 2001.
No. As applied here to computer code, the anti-trafficking provisions (specifically §1201(a)(2) and (b)) are content-based in that they specifically target the publication of, and linking to, scientific expression based on the particular topic addressed by that expression -— namely, techniques for circumventing CSS. The fact that the expression here was conveyed in a professional language, computer code, does not change the content-based nature of the provisions. The restriction at issue here is no different from a regulation prohibiting engineers from using blueprints, as opposed to English prose, to explain how to build a photocopier, without applying the same prohibition to discussion of other kinds of machines.
But even if §1201(a)(2) and (b) could be construed as content neutral, they would still need to meet the stringent requirements of Turner Broadcasting System, Inc. v. Federal Comm. Commission, 512 U.S. 622, 663 (1994)("Turner I"), standards they cannot meet as described infra at 13-15 and 17-20.
Moreover, even if the statute could be construed as content neutral, the District Court's injunction here should not be. The injunction targets only speech – the publication of and linking to DeCSS . It specifically bars "posting on any Internet web site" or any link to another web site "for the purpose of disseminating DeCSS." Universal, 111 F. Supp. 2d at 346-347 (amended final judgment), and therefore should be receive strictest scrutiny.
And even if the injunction could be construed as content neutral, the Supreme Court recently reminded us that the application of such relief to a member of the media requires scrutiny "of the highest order." Bartnicki v. Vopper 532 U.S. ___ (May 21, 2001) (slip op. at 12) (citing Smith J. Daily Mail Publishing Co., 443 U.S. 97, 102 (1979)). In Bartnicki, the Supreme Court applied this high level of scrutiny to a content-neutral law forbidding the "disclosure" of a tape of an illegally intercepted cellphone conversation when the information contained was a matter of public concern and the defendants did not participate in the illegal interception. The tape concerned a teachers' union negotiation and was broadcast by media. Here, where § 1201 has been applied to forbid media publication of the information contained in a computer program that is a matter of public concern, the First Amendment requirements emphasized in Bartnicki are squarely raised.
No. DeCSS itself has no non-speech elements. It is a set of instructions written in a specific professional language that expresses ideas to those who can read that language. Computer programmers and scientists communicate using programming languages because these languages are an unambiguous mode of expression. Universal, 111 F.Supp.2d at 345, n. 225 (testimony of Dr.Touretzky). These facts and the legal cases supporting them are further described in the Amicus Brief of Abelson, et al. at 13-17.
That a person might use a computer program to do something does not by itself add "nonspeech" elements to the text. Blueprints and instructions for a photocopier, recipes, books about fixing cars, and videos on baby care surely count as fully protected speech, no matter how "functional" their content might prove to be. Prohibition of these works merely because recipients might easily do what they describe or explain would require searching First Amendment review and justification.
The fact that DeCSS is pure speech does not, of course, prevent regulation of computer programs any more than it prevents regulation of other kinds of dangerous speech. Settled laws of aiding and abetting liability and conspiracy allow prosecution of those who disseminate computer programs (see infra at ___). However, the mere fact that programs can be used to direct the functioning of computers, by itself, does not require the creation of a new category of "non-speech" or partial "non-speech" for them.
Moreover, even when the "functionality" of DeCSS is considered, it is important to remember the limited function of this particular program. Despite the Government's inflammatory rhetoric, DeCSS is not intrinsically harmful. No one can use DeCSS to "shut off navigational systems on airplanes or shut down smoke detectors in public buildings." By itself, DeCSS does not even perform any infringing activity. A person can use DeCSS to do only one thing -- access the content on a DVD. From that point a person must choose what to do with the decrypted movie. She could play it on a Linux computer, copy a snippet for a book report, make a personal back-up copy, or feed it into a database to be used for scholarly research, all permissible, fair uses of the DVD content. See e.g. Universal, 111 F.Supp.2d at 322.
More disturbing than its factual inaccuracy is the disregard for the First Amendment in the Government's rhetoric. It argues against the free exchange of scientific and technical information through the press on matters of public concern by making irresponsible comparisons between the dissemination of such information and acts of deliberate, criminal harm by those who receive the information. It is one thing to prosecute those who would jam aircraft navigation, but quite another to prohibit engineers from discussing technologies that, in the hands of unknown criminals, might be used for this purpose. This claim demonstrates why, as the Supreme Court has just reminded us in Bartnicki, the courts must be vigilant in their application of First Amendment standards.
No. The dissemination of DeCSS, here by a member of the media covering an issue of public concern, is pure speech. See e.g. Bartnicki at 11-12 ("It is true that the delivery of a tape recording might be regarded as conduct, but given that the purpose of such a delivery is to provide the recipients with the text of recorded statements, it is like the delivery of a handbill or a pamphlet, and as such, it is the kind of "speech" that the First Amendment protects.") That the communication occurred on the Internet does not change the analysis. The Internet is "a unique and wholly new medium of worldwide human communication" that receives full First Amendment protection. Reno v. ACLU, 521 U.S. 844, 850, 879 (1997).
It can. The use of DeCSS always facilitates access to expression, since the use of DeCSS to decrypt an encrypted DVD simply allows access to a movie, just as a newsrack allows access to a newspaper, a CD player allows access to the music on a CD and a word processor allows access to a document. Such acts are analyzed as activities "commonly associated with expression" for which First Amendment scrutiny applies. See City of Lakewood v. Plain Dealer Publishing, 486 U.S. 750, 769 (1988). Use of DeCSS to decrypt can have similar expression-related uses, such as allowing the viewing of a movie one has purchased, fair use of DVD content, or scientific analysis of the encryption programs used on DVDs.
But the use of items that give access to speech may not be fully protected by the First Amendment. A sound truck or newsrack has no such protection when used to block the sidewalk or street, nor does a word processor when used to create a ransom note. Similarly, the use of DeCSS to aid in copyright infringement is not "commonly associated" with protected expression, and thus may be prosecuted as copyright infringement without running afoul of the First Amendment.
No. Many speech activities, like flag-burning and charitable solicitation, have non-speech elements. In both situations, regulation that impacts the speech elements based upon their content merits "the most exacting scrutiny." United States v. Eichman, 496 U.S. 310, 318 (1990) (flag burning); see e.g. Village of Schaumburg v. Citizens for a Better Environment, 444 U.S. 620, 632 (charitable solicitation). In other words, it is not the presence or absence of non-speech elements, but rather the content-based nature of the challenged regulation, that determines the appropriate First Amendment scrutiny.
No. There is no First Amendment precedent establishing that a "causal link" between publication of speech and its "improper use" justifies the imposition of intermediate scrutiny. As the Supreme Court recently observed, "it would be quite remarkable to hold that speech by a law abiding possessor of information can be suppressed in order to deter conduct by a non-law-abiding third party." Bartnicki at 14. The logic behind §1201 is the same -- suppressing the speech of 2600 Magazine in order to deter copyright infringement by non-law-abiding third parties. And here, not only was the information lawfully acquired, it was publicly available in many other places prior to its publication by 2600 Magazine and is still widely available today.
Even when the speech has a close causal link to violent activity, as was considered in Brandenberg, National Association of Colored Persons v. Claiborne Hardware, 458 U.S. 886 (1982), and most recently in Planned Parenthood v American Coalition of Life Activists, 244 F.3d 1007 (9th Cir. 2001); 2001 WL 293260, courts have consistently rejected the argument that a mere "link" between speech and illegal activity is sufficient grounds to prevent or punish the speech.
Instead, speaker liability is not allowed "unless that speech is capable of producing imminent lawless action." Brandenberg at 447. As the Planned Parenthood court noted, "[I]t doesn't matter if the speech makes future violence more likely; advocating 'illegal action at some indefinite future time' is protected." Planned Parenthood 244 F.3d at 1015 (citing Hess v. Indiana, 414 U.S. 105, 108 (1973) (per curiam)).
If the First Amendment does not permit imposition of liability upon a pro-life group for publishing individuals' names and addresses, a publication that caused the effected individuals to take extraordinary steps to protect themselves and their families, it should not permit imposition of liability upon 2600 Magazine based merely on the assertion that DeCSS might increase the likelihood that plaintiffs' copyrighted DVD movies will be infringed by third parties.
Outside the context of the alleged "advocacy" of violence, the law still consistently requires more than a mere "causal connection" in order to impose liability on a speaker based upon the acts of the recipients of his speech. Aiding-and-abetting liability, which reaches speech, has two related requirements, neither of which exists in §1201. First, it requires intentional wrongdoing, not mere recklessness, by the abettor. Central Bank of Denver N.A. v. First Interstate Bank of Denver. N.A., 511 U.S. 164, 190 (1994); see 18 U.S.C. § 2. Focusing on intentional wrongdoing creates breathing space for publishers like the New York Times and 2600 Magazine.
Second, aiding-and-abetting liability requires that the underlying offense in fact be committed, although the government need not prove by whom it was committed. See e.g., United States v. Superior Growers Supply, Inc., 982 F.2d 173, 177-78 (6th Cir. 1992); United States v. Campa, 679 F.2d 1006, 1013 (1st Cir. 1982). Here, of course, no acts of copyright infringement have been linked to 2600 Magazine's publication of DeCSS.
Similarly, speech that is part of a conspiracy can only be punished if it is shown the speaker (i) knew of the intended unlawful use of the information and (ii) agreed with other conspirators that an offense would be committed. See generally Direct Sales Co. v. United States, 319 U.S. 703 (1943); United States v. Pinckney, 85 F.3d 4, 8 (2d Cir. 1996). § 1201 does not require either element.
Finally, even under intermediate scrutiny, the government bears the burden to establish the closeness of a causal link between the publication of a text that can have both legal and illegal purposes and the illegal activity allegedly facilitated by the use of the text. See e.g. Turner I at 663 (government must do more than posit the existence of a disease to be cured). Neither an unsubstantiated legislative pronouncement nor a judicial hunch -- and certainly not the unproven "actual threat" of illegal activity claimed by the Appellees at oral argument -- is sufficient. Intermediate scrutiny demands substantial evidence, independently reviewed by the court, for determining the close causal link between the dissemination of DeCSS and any allegedly imminent illegal use.
No. As noted above, Brandenburg Court held, "the constitutional guarantees of free speech and free press do not permit a State to forbid or proscribe advocacy of the use of force or of law violation except where such advocacy is (1) directed to inciting or producing imminent lawless action and (2) is likely to incite or produce such action." Brandenberg, 395 U.S. at 448. Neither § 1201 on its face nor the District Court's application of it required proof of intent, imminence or likelihood of illegal action; mere illusion to "substantial risk" falls far short of Brandenberg's strict requirements. At a minimum, a remand would be required to permit the court to make findings on these questions.
No. We note that the use of DeCSS has not been enjoined; the injunction targets only its dissemination. We further note that, as discussed above, the Supreme Court has recently affirmed that more searching review is required where media publication of lawfully acquired information about a matter of public concern is at issue. Bartnicki at 12.
Intermediate scrutiny requires that the regulation of speech must "not burden substantially more speech than necessary" to advance the government's interests. Turner I, 512 U.S. at 661; Turner Broadcasting v. FCC, 520 U.S. 180, 185 (1997). Under the Turner standard, the burden rests upon the government to "demonstrate that the recited harms are real, not merely conjectural, and that the regulation will in fact alleviate these harms in a direct and material way." Turner I, 512 U.S. at 664. Such a showing requires an empirical record "assur[ing] that, in formulating its judgments, Congress has drawn reasonable inferences based on substantial evidence." Id. at 666, and "judicial findings concerning the availability and efficacy of ‘constitutionally acceptable less restrictive means' of achieving the Government's asserted interests." Id. at 668.
The District Court's three criteria obviously fail these requirements. First, by allowing a near total ban on the dissemination of circumvention tools they ignore the requirement that the statute "burden no more speech than necessary." The criteria are completely silent as to circumvention of access controls for legitimate, speech-related purposes, including fair use and scholarly study of encryption and circumvention methods.
Second, as noted above, Turner requires searching consideration, by both Congress and a reviewing court, of the availability of less restrictive means, not merely a finding by the court that there is no other practical means of preventing infringement than banning the publication. As explained in further detail below, Congress had many less restrictive options that it failed to consider in any material way. See infra at 17-20.
Third, the "governmental motivation" criterion adopted by the district court conditions liability on how a program "may be used" instead of limiting it to how the program has been used. Moreover, it ignores the innate nature of computer programs by asserting a specious distinction between the "function for which the programs exist" and the "message they may convey." As Dr. Touretsky and the Abelson amici demonstrated, what computer programs say cannot be divorced from what they permit people to do with them. Thus, the District Court's test expands the injunction standard from one targeting illegal behavior to one that inevitably requires changes in the content and message of the program itself. And it does so even before any harm has occurred.
As explained both above and below, Turner first requires consideration of whether §1201 itself is constitutional, including consideration of whether Congress had less restrictive alternatives. But even if §1201 is facially justified, it may not be constitutionally applied to publication of a program such as DeCSS unless the District Court has found:
1. That there is no noninfringing use, fair use or free speech right to offer DeCSS.
2. That a sufficient relationship exist between 2600 Magazine and infringers under the aiding and abetting or conspiracy standards; and
3. That DeCSS was used to infringe copyrights or that an imminent danger of copyright infringement exists under the standards announced in Brandenburg.
If these modifications cannot be made without diverging extensively from the legislative framework, the statute must be struck on its face.
No. As government acknowledged at oral argument, the District Court's announced test for the injunction of links could bar even the publication in a newspaper of a list of addresses where contraband books could be found. It would require a court to evaluate the motives behind the publication of challenged links in deciding whether to impose § 1201 liability, and that such a searching evaluation would apply as much to links published by the New York Times or Professor Jane Ginsburg (see Appellants' Opening Brief at 14) as those published by 2600 Magazine. The chilling effect on speech caused by the specter of such ex post facto evaluation of a publisher's motives is plain.
Once again, the recent Bartnicki decision addresses the infirmities of this test. In Bartnicki the statute was held to be insufficiently protective of speech when it created liability for the willful disclosure of the contents of a wire communication with knowledge that the information was obtained through illegal intervention. Bartnicki at 8. Here, similarly, the District Court's test requires only the willful linking to DeCSS with the knowledge that it may not lawfully be offered and the intent to disseminate -- not the intent to cause or aid and abet any harm. Bartnicki teaches us yet again that restraining the press from conducting its legal activities in the interest of limiting others' wrongdoing is not only bad social policy: it is also forbidden by the First Amendment.
A link is simply a reference, indicating that certain information is available in a particular place. A link should be enjoined when, and only when, the court would also enjoin a similar reference by a print publication. Even supposing, arguendo, that DeCSS is speech not covered by the First Amendment, this injunction should be affirmed only if, to take an example suggested by the Court at oral argument, it would be constitutional to enjoin a newspaper from giving the address of a bookstore where allegedly obscene books may be found in the course of reporting on an ongoing debate about adult booksellers. The comparison, even when the subject is not protected speech, shows how far beyond the boundaries of accepted First Amendment doctrine the District Court strayed here.
As noted above, even before the specific application of the statute is addressed, Turner requires a judicial finding that Congress properly considered less restrictive means to meet its goal. Here, Congress enacted a near total ban on the dissemination of circumvention technology, so the inquiry must look at whether this drastic step was necessary. In doing so, this Court "can take Congress' different, and significantly less restrictive, treatment of a highly similar problem as at least some indication that more restrictive means are not ‘essential' (or will not prove very helpful)." Denver Area Educational Telecommunications Consortium v. FCC, 518 U.S. 727, 757 (1996). Here, any of the alternatives listed below would suffice to defeat the Appellees' claim.
a) Congress could have restricted circumvention liability to those who have intentionally aided and abetted copyright infringement or who have engaged in a conspiracy to infringe copyrights, following the path taken by "burglars tools" statutes. See supra, question 8; e.g. Appellants' Reply Brief at 17-19; ACLU amicus brief at 19-22.
b) Congress could have limited liability to circumvention for illegal purposes, as outlined in the Boucher and Ashcroft bills, rather than extend liability to tools makers and tools disseminators. H.R. 3048, 105th Cong. § 8 (1997) (Boucher); S. 1146, 105th Cong. tit. III (1997) (Ashcroft).
c) Congress could have protected fair and noninfringing uses the same way it protected the ephemeral recordings in the DMCA. 17 U.S.C. §112(a)(2). There, Congress expressly required copyright owners to "make available to the transmitting organization the necessary means" for making ephemeral copies and authorized self-help if the means are not made available "in a timely manner." Id. Congress could have done the same for fair use: required that copyright owners utilizing technological protection measures make copies of their work available for fair or noninfringing uses, allowing circumvention and the dissemination of the required circumvention programs should they fail to do so.
d) Congress could have followed the model of the Vessel Hull Design Protection Act, 17 U.S.C. §1309, passed as part of the same legislative package that contained § 1201. In that statute, Congress required that a disseminator of information be liable only if he or she "induced or acted in collusion with" one who actually gains unauthorized access to a work.
e) Congress could have followed the model it used for analog VCRs. Although 17 U.S.C. §1701(k) requires that VCRs conform to copy control technologies, it also generally prohibits the use of such technologies "to prevent or limit consumer copying." 17 U.S.C. §1201(k)(2). Congress provided that failure to allow consumers to copy is itself an unlawful "act of circumvention" by the content holder. 17 U.S.C. §1201(k)(5).
f) Congress could have followed the model of the Audio Home Recording Act ("AHRA"), 17 U.S.C § 1002(a), which permits fair and noninfringing users to make one copy of a copyrighted work but allowed the prevention of serial copying.
Each of these steps would have been less restrictive than § 1201. Each would have been as effective. They therefore demonstrate that Congress failed to adequately consider the balance between speech and copyright protection. In fact, Congress' approach turned that balance on its head. Normally, speech is protected and potential harms to others are remedied as they arise. In contrast, in passing §1201, Congress barred all nonconsensual access to digital expression in an effort to prevent economic harm to copyright owners that may someday arise, while leaving speech to be protected by later legislation or caselaw.
At oral argument, the Court asked whether Congress' failure to consider the obviously less restrictive means of the AHRA was due to a lack of applicable technology. The answer is no. Sophisticated copy control mechanisms for DVDs that limit users to a limited number of copies, which copies cannot be further re-copied, exist and were actively being developed even before §1201 was enacted. Thus, the assertion of the government and Appellees during argument that a "single copy" limitation such as that provided by AHRA would not be sufficient protection today because that single copy could be "uploaded" to the Internet and thereafter copied by an unlimited number of others is simply wrong.
At the same time that the Appellees were seeking §1201 from Congress, they were also seeking proposals for an AHRA-type system for DVDs. In July, 1997, the DVD Copy Protection Technical Working Group ("DVD CPTWG"), a coalition made up of the Appellees and consumer electronics manufacturers, issued a call for proposals that included in its requirements serial copy management capability of "generational control for one copy," including three "states" in which content could be distributed: "never copy," "no more copy" and "one copy."
Serial copyright management schemes can be set to allow any set number of copies and to follow a number of rules, including allowing the copying of a number of certain sized snippets from a work. For example, in 1997 Digimarc described the technology it had developed with Macrovision as follows: "[D]igital watermarks embedded in movies and other video material act as an invisible identity providing . . . rules to either allow or disallow playback, viewing, and copying onto another digital recorder."
Beginning September, 1997, the DVD CPTWG received eleven proposals in response to its request from such companies as IBM-NEC, Pioneer-Hitachi-Sony, Macrovision-Digimarc-Phillips. Thus, Congress' failure to consider an AHRA model for DVDs when it enacted §1201 in 1998 was not because DVDs could not be distributed with serial copyright management schemes.
While serial copy management technologies do not enable the full range of uses allowed under the fair use doctrine or fully preserve other noninfringing uses, these systems are certainly less restrictive than the complete prohibition on access to CSS-protected works created by the District Court's interpretation of §1201. Coupled with a proper statutory model, these technologies could provide protection against copyright infringement without the severe impact on free speech, fair use and the public domain created by §1201.
At argument, the Court asked whether "fair use" meant a right to access the work in the technologically most usable form. The short answer is that fair use extends to works in whatever form they are offered to the public.
The longer answer is that, since fair use is the safety valve by which the First Amendment and copyright exist peacefully, if different media permit different levels of individual expression through fair use, then decisions by Congress to impede the most useful means require justification under the First Amendment. A prohibition on using copy machines to make fair use could not be answered -— under today's First Amendment law —- with the retort that one can hire monks to scribe the relevant passages. Similarly, the possibility of disseminating indecent materials by videotape could not save a prohibition on the distribution of such materials via the Internet. See Reno, supra. The claim that a lesser means of communications would be available to users no more answers the First Amendment question here than in these other cases.
No one can force the Appellees to release their works in digital form. Once they do so, however, the First Amendment underpinnings of fair use and the limited nature of the copyright holder's exclusive rights require that these limits on copyright apply unless the Government can sustain its heavy burden to show why these expressive uses should be banned in this new medium of expression.
By enacting the DMCA, Congress intervened early at the behest of powerful commercial interests to regulate a robust and expanding market for digital expression. It did so by abandoning the traditional balance of copyright law and by giving the Appellees vastly greater control over access to their works than copyright law has traditionally provided. See Netanel, Locating Copyright Within the First Amendment Skein, 54 Stanford Law Rev. 63 (May 2001 Draft) (forthcoming October, 2001). According to the government at argument, the reason was as much about providing the Appellees a subsidy to induce them to go online as it was about a concern that changing technology would undermine copyright protection. But burdening fair and noninfringing use—the privileged expression of users—to subsidize preferred speakers is constitutionally infirm.
We do not question the right of Appellees to protect their works. Nor do we argue that they must publish their works in digital form. We only claim that when Appellees choose to publish copies of their works in digital form, those copies are subject to limiting principles, some of which derive from the Constitution. Even if code may protect works from copying more completely than the law, the law may reinforce that code only to the extent the Constitution allows.
Attorneys for Appellants