Copyright Office's Draft of S 2560, the "Inducing Infringement of Copyrights Act of 2004", with Explanatory Memorandum.
Date: September 9, 2004.

Editor's notes:
 • TLJ created this document by converting an MS Word document into HTML.
 • TLJ changed fonts, spacing, and paragraph indentations.
 • TLJ added hyperlinks.

Recommended Statutory Language

Section 501 of title 17, United States Code, is amended by adding at the end the following:

(g)(1) Inducement of Infringement -- Whoever manufactures, offers to the public, provides, or otherwise traffics in any product or service, such as a computer program, technology, device or component, that is a cause of individuals engaging in infringing public dissemination of copyrighted works shall be liable as an infringer where such activity:

(2) For the purposes of this subsection, “public dissemination” means digital transmission to the public of copies or phonorecords or any other exercise of any of the rights set forth in sections 106(3), 106(4), 106(5) or 106(6).

(3) Limitations on Remedies -- (A) No award of statutory damages under Section 504(c) shall be made for a violation of this subsection unless the copyright owner sustains the burden of proving, and the court finds, that such violation was committed willfully. (B) In granting injunctive relief under Section 502 for a violation of this subsection, the court shall, to the extent practicable, limit the scope of the injunctive relief so as not to prevent or restrain noninfringing uses of the product or service.

(4) Nothing in this subsection shall enlarge or diminish the doctrines of vicarious liability and contributory infringement, including any defenses thereto or any limitations on rights or remedies for infringement.  Nothing in this subsection shall enlarge or diminish liability for infringement of the exclusive rights in sections 106(1) or 106(2).

Explanatory Memorandum


In developing this recommendation, the Copyright Office reviewed comments from and engaged in constructive discussions with a wide variety of interested parties, including American Federation of Television and Radio Artists, Association of American Publishers, Broadcast Music, Inc., Business Software Alliance, Center for Democracy & Technology,  Consumer Electronics Association, Corbis, Digital Future Coalition, Digital Media Association, Distributed Computing Industry Association, Entertainment Software Association, Future of Music Coalition, IBM, IEEE, Information Technology Association of America, Internet Commerce Coalition, Motion Picture Association of America, NetCoalition, Public Knowledge, Recording Artists Coalition, Recording Industry Association of America, Software and Information Industry Association, Sun Microsystems, Verizon, and the Video Software Dealers Association. Senate Judiciary staff met with several groups individually to discuss the bill and examine proposed changes, and they shared the results of those meeting with us. We reviewed the thoughtful proposals for statutory language submitted by IEEE, NetCoalition, Consumer Electronics Association and the Business Software Alliance.

After the individual meetings held by Committee staff, Jule L. Sigall, Associate Register for Policy & International Affairs, moderated a group meeting in late August with interested parties identified by the staff.  At this meeting, the various alternatives and ideas contained in the proposals were discussed with the parties in a constructive dialogue.  Based on that discussion and the written proposals, the Copyright Office circulated a discussion draft to the parties that was intended to reflect the ideas discussed in the meeting in more concrete terms.  We made clear that this discussion draft was not a draft of the Office’s recommendation; rather, it was designed to provoke comment and feedback on the many different concepts put forth in the various proposals so that the Office could be fully informed on the parties’ views on all aspects of this issue.

We received written comments from nearly all of the interested parties on the discussion draft, and held a meeting on September 7 to discuss those comments with the parties.  Several groups who commented on the discussion draft but were not originally identified by Committee staff as participants requested to and did participate in that meeting.  In all cases, these meetings and discussions were highly constructive, with the parties considering the issues forthrightly and providing thoughtful feedback and information on the many different issues involved.

Description of Recommendation

This recommendation reflects an effort to meet the goals expressed for this legislation by the cosponsors of S. 2560 at the July 22 hearing and in their August 13 letter.  First, the bill should be technology-neutral, and should not impose liability based on the use of any particular technology but on the circumstances in which that technology is used.  Second, it should provide an effective cause of action against those services who today are establishing massive networks of copyright infringement through the use of peer-to-peer technology, but be flexible enough to accommodate new, unforeseen situations in the future.  Third, the bill should not chill innovation by unnecessarily creating the threat of liability for developers and distributors of technology who are not engaged in encouraging and profiting from copyright infringement.  Fourth, the bill should also preserve the decision by the Supreme Court in Sony Corp. v. Universal City Studios, Inc., 464 U.S. 417 (1984), often referred to as the Betamax case.  We have strived to meet all of these goals in our recommendation, despite the great difficulty that exists in reconciling these important and sometimes conflicting issues. 

In light of the variety of conflicting proposals put forth by the parties and the many issues involved, it was not possible to achieve consensus for this recommendation in the time we had.  What we have attempted to do is recommend an approach that accommodates the legitimate concerns of all the parties, and therefore we believe it presents the best chance to serve as basis for developing consensus for this legislation.

Our effort to develop consensus proposals began with the text of S. 2560, and we engaged in a discussion with the interested parties of additional language that would clarify the scope of the bill and assuage concerns of technology companies and others about its application.  The bill as introduced reflects a “behavior-based” approach, which has the decided advantage of focusing not on the technology involved but on the defendant’s actions.  NetCoalition and IEEE offered alternatives that would codify all secondary liability doctrines, including a form of inducement liability.  Another approach was offered by Consumer Electronics Association, which would create liability for a narrow class of activity of distributing a computer program specifically designed to encourage mass, indiscriminate infringing distribution.   The Business Software Alliance proposed changes to S. 2560 that it asserted would make the bill more specific and give more comfort to technology companies about the scope of liability under the bill.  Our discussion draft attempted to capture as many of the concepts in these proposals as possible so that we could hear the parties’ views on them before making our recommendation.

A concern raised by technology companies about the behavior-based approach was that it did not provide enough guidance on what activity was subject to liability and what was not.   Also, it was claimed that an approach that focused on intent would likely not be susceptible to summary adjudication, but would require expensive jury trials for resolution of the liability question.  To address these concerns, some parties offered the CEA proposal as an alternative, claiming that it offered more objectivity and a clearer standard of liability, even though it, by its terms, contains elements of intent and behavior, and an element examining the design of technology.

Our discussion draft provoked many comments, and parties on both sides had concerns about many of the provisions.  One concept that did not draw much criticism, and garnered some support from one technology group, was a provision that made liability turn on whether the defendant provided technology as part of an operation that relied on infringement for its commercial viability or for a predominant portion of its revenues.  The CEA proposal also contains a variation of this concept, albeit used in a different context.

After serious consideration of the issues raised by these alternative approaches, we have concluded that none of these approaches by itself accommodates the legitimate concerns of all parties and meets the goals described above.  However, each has positive aspects to it that have contributed to our recommendation.  We have concluded that enacting a separate form of liability, as S. 2560 would, is the preferable course to take, rather than codifying all forms of secondary liability, given the long history these existing doctrines have and the myriad of additional issues that would be raised if their codification were considered as part of this effort.  We started with the general concept in the CEA proposal of clearly identifying the activity that creates liability, but modified that draft substantially to make it meet the goals described above, particularly the goal of creating a form of liability that can accommodate changes in the way technology is used in the future. 

This approach focuses neither on technology nor behavior generally, but specifically on the business model in which a technology is employed, essentially revolving around the question of whether the defendant is relying on infringement to build a business around its product or service.  Our decision was based in part on the relatively positive feedback described above that we received on this “business model” concept.  We believe that this approach best serves the goal of both capturing the existing peer-to-peer services built on widespread infringement, yet remaining flexible enough to address potential use of new technology by “bad actors” in the future that we cannot predict right now, just as we could not predict the imminent phenomenon of massive infringement by means of peer-to-peer when the Digital Millennium Copyright Act was enacted in 1998.  Experience has taught us that legislation that focuses on a particular technology is likely to be obsolescent from the moment it is enacted, and waiting to enact new legislation to address new technological developments inevitably is a Sisyphysian effort.

Subsection 501(g)(1)

Our recommendation would create a new subsection (g) to section 501 of the Copyright Act.  The operative provision on liability is set forth in paragraph (g)(1).  This provision narrows the existing bill considerably, as it only applies to situations where technology is employed as a cause of “infringing public dissemination.”  As explained below, “public dissemination” is defined in paragraph (2) to exclude direct infringement that implicates only the reproduction or derivative work rights. In other words, this recommendation would not impose liability on any product or service that is a cause of only private reproductions, such as the VCR, CD-burners, or the typical portable music player.  Liability, if any, for conduct related to such personal reproduction technology remains the province of existing copyright law, and is not affected in any way by this new form of liability.  Furthermore, liability would be imposed only in one of three circumstances, each of which turns on the importance of the infringing use of the product or service to the defendant’s commercial interest or efforts to attract the public to their product or service.  Liability would result if the technology is “a cause” of individuals engaging in infringing public dissemination of copyrighted works. We considered  predicating liability on whether the defendant has “induced” or “caused” infringement, but after careful consideration concluded that neither of those words clearly expresses the appropriate causal connection between the actions of the defendant and the infringing acts of a third party. Although the theory of liability is based on concepts of inducement, as reflected in the caption of subsection (g)(1), discussions on this issue revealed a great deal of concern over the precise meaning of the term “induce” and little agreement over what the meaning of that term as used in the bill should be. Predicating liability on whether the defendant “causes” infringement could be interpreted as requiring that the defendant’s conduct be the proximate or ultimate cause of the infringement, which we rejected as too high a burden. Obviously, the proximate cause of any act of infringement is most likely going to be the volitional act of the individual who engaged in the infringement, rather than any acts of a person who has induced an individual to infringe.  We concluded that, combined with the three conditions set forth in subparagraphs (g)(1)(A)-(C), simply requiring that the defendant’s conduct be “a cause” of the infringement reflects the appropriate balance.

The key factor in determining  liability lies in the three conditions, at least one of which must be met, in order for liability to be imposed.  Each of these conditions set forth in subparagraphs (g)(1)(A)-(C) turns on the importance of the infringing use of the product or service to the defendant’s commercial interest or efforts to attract the public to its product or service.  The third condition in subparagraph (C) would allow liability to be imposed where a defendant attracts consumers to its product or service primarily through infringement but has not undertaken the activity for commercial purposes.

The approach taken here seeks to avoid the controversial issues surrounding the various alternatives and focus liability on the business model adopted by the defendant.  If the defendant distributes its product or service as part of an enterprise that relies on infringement to attract customers, remain commercially viable or earn a predominant portion of its revenues, it will be liable under this provision.  Under this approach, there is no need for the courts to get into thorny questions about how a particular technology was designed, what its particular functions can or can’t do, or difficult questions of intent.  In essence, this approach determines whether inducement has occurred through an objective test of how the technology is deployed as part of the defendant’s business. 

The analysis under these three provisions would focus on the defendant’s “activity” of manufacturing, distributing or providing the product or service, and in cases where a particular activity is a part of a larger operation, the court should isolate the defendant’s activity that is a cause of the infringement at issue from the rest of the organization for purpose of analyzing whether any of the three conditions apply.  For example, if a large software and services company has a division that begins distributing a device that is a cause of infringing dissemination, the revenues and commercial viability of the other parts of the company would not be considered in the analysis of subparagraphs (g)(1)(A)-(C); thus, whether or not an entire organization’s revenues are derived from infringement would have no bearing on liability, but only whether the activity in question depends on infringement for its commercial viability or that activity’s revenues predominantly derive from infringement.  We intend “derives a predominant portion of its revenues from infringing public dissemination” to mean that more of the revenues from that particular activity can trace their origin to infringing public dissemination (e.g. to purchases by persons who are using the product or service for purposes of infringing public dissemination) that to all other sources; i.e. more than fifty percent of revenues are derived from infringing public dissemination. 

Section 501(g)(2)

Paragraph (2) defines “public dissemination” as the exercise of the so-called “public” rights under copyright:  distribution of copies to the public, public performance, public display and public performance of sound recordings by means of digital audio transmission.  It also makes clear that distribution by digital transmission to the public is covered, to leave no doubt that the form of infringing distribution occurring on peer-to-peer services falls within the form of direct infringement that serves as a predicate for liability under this new provision.  The term “digital transmission” is already defined in section 101 of the statute, and a similar formulation to that here is used in Section 115(c)(3)(A), which makes clear that the mechanical compulsory license includes the right to “distribute or authorize the distribution of a phonorecord … by means of a digital transmission ….”  See also, e.g., A&M Records, Inc. v. Napster, Inc., 239 F.3d 1004, 1014 (9th Cir. 2001).  We have also included the public performance and public display rights to cover services that do not distribute copies but still infringe by making works available to the public in the form of performances and displays, such as peer-to-peer webcasting services.  However, the mere infringement of the reproduction right or the derivative work right, without any of these additional elements, would not constitute “public dissemination.”  This should allay the concerns of those who fear that the legislation could be used to target manufacturers and marketers of devices used for personal copying.

Section 501(g)(3)

Because paragraph (1) makes those liable under this subsection as “an infringer,” all of the remedies available under Chapter 5 of the Copyright Act are available in actions under this subsection.  However, paragraph (3) provides two important limitations on those remedies, one of which is based on a proposal made by IEEE regarding limitations on monetary relief.  First, statutory damages, of any level, are not available for violations of this new provision unless the copyright owner proves that the violation of subsection (g)(1) was willful.  This provision was drafted to address concerns raised by technology companies that the specter of liability under this theory might chill innovation, and a limitation on statutory damages should help to ally those concerns, since a nonwillful violator would only be liable for the actual damages and lost profits that the plaintiff could prove the defendant had caused.  Also responsive to this concern is the existing Section 505 of the Copyright Act, which allows a court to award attorney’s fees to a prevailing defendant in “any civil action” under title 17, which would include this new form of liability.

Second, a court that issues an injunction under this subsection must, where practicable, fashion the scope of the injunction not to restrain or prevent the noninfringing uses of the product or service at issue.  This is a critical aspect of the Sony decision, which sought to prevent copyright liability from inhibiting the noninfringing use of technology.  Thus, in the peer-to-peer context, an injunction might prohibit those features of a peer-to-peer service that cause infringement, but would not prohibit the service or the technology itself.

Section 501(g)(4)

This paragraph provides savings clauses to ensure that this new form of liability does not affect the existing doctrines of secondary liability, as well as the defenses and limitations on remedies for such liability.  It also makes clear that this new provision does not affect existing law of liability for infringement of the rights of reproduction and preparing derivative works, which are not predicate acts for liability under this recommendation.

As with the remedies provisions in Chapter 5, which are fully applicable to this new cause of action, the limitations on liability in Section 512 also apply fully to this new provision.  Thus, even in the unlikely case that a qualifying “service provider” would meet the criteria for liability under this provision, it would still be entitled to the limitations on liability in Section 512, if it met the conditions of those provisions.

Like the new form of liability contained in the CEA proposal, the recommendation does not make the defense recognized in the Sony decision regarding capability of noninfringing uses applicable to this new form of liability.  Application of the Sony defense to this new form of liability is unnecessary, as the concerns underlying that decision are accommodated in this recommendation.  First, the recommendation’s narrowed scope, addressing technologies of “public dissemination,” confines it to circumstances not present in the Sony case, which involved personal copying technology in the form of the VCR, not distribution technology.  Second, liability under this provision does not turn on a product’s or service’s capabilities, features or design – rather, it depends on the business model in which that product or service is offered.  The very same technology may result in liability when offered by one defendant and not result in liability when offered by another, depending on how each defendant relies on the infringing use of the technology to attract consumers or earn revenues.  In this way, the recommendation avoids the concern expressed by groups like IEEE and BSA over having the courts engage in searching review of a technology’s design or capabilities to determine liability.  Third, as noted above, the limitation on injunctive relief that requires a court to accommodate noninfringing uses as much as practicable vindicates the most important concern underlying the Sony case – that unrelated areas of commerce not be burdened by copyright liability.  Fourth, the savings clause makes clear that Sony remains fully applicable to causes of action under the existing doctrines of secondary liability.