2002 UCLA J.L. & Tech. 5

The Digital Evolution:
Freenet and the Future of Copyright on the Internet
by Ryan Roemer

Introduction

Peer-to-peer ("P2P") technologies have created more criticism and praise than perhaps any other recent Internet technology. Working without the aid of a central server, P2P data sharing brings the Internet back to its roots, allowing users to connect directly with each other and exchange files. Unfortunately, bypassing these intermediary servers has left a legal system that assumes a central responsible party struggling to adjust.

In 1999, a college student named Shawn Fanning created a music file-sharing program called Napster. Napster almost immediately brought millions of users together, downloading and sharing an unprecedented amount of MP31 files, most of which were copyrighted musical works.2 By December of 1999, as Napster's popularity gained, the record industry brought suit against Napster, Inc. for copyright infringement.3 Although the recording industry eventually prevailed on their copyright claims and largely killed off most of Napster's user base, the damage had already been done. File-hungry users quickly flocked to the next generation of P2P technologies, which are adapting to new legal and technological hurdles at an amazing rate. The recording industry has since launched legal attacks on a variety of fronts, going after new P2P companies as well as individual users.4

This paper examines a developing technology named Freenet.5 Freenet is a P2P architecture designed to be secure, efficient, and built to withstand virtually any legal or technological challenge. Although it is far too soon to predict Freenet's triumph or defeat, it is a worthwhile conceptual model to examine, because modern P2P systems are already integrating many law-defying characteristics of Freenet and will continue to do so, even if the Freenet project never takes off.

The thesis of this paper is that the technology of P2P systems is stretching the applicability of modern copyright law to its limits. Content owners will possibly see the day that a technology like Freenet will render the current law ineffective against digital copyright infringement on the Internet. However, it is more likely that both P2P technology and copyright law will ultimately survive the current conflict intact, as they both adapt to legal and technological changes driven by worried content owners, digital libertarian programmers, and the file-hungry public.

Part I of this paper examines the technological nuances of Freenet and the trends in P2P projects to develop systems that withstand both legal and technological attack. Part II documents the ongoing battle between content owners and companies touting new, disruptive technologies that threaten copyrights, ranging from video cassettes to file-sharing computer programs. Part III examines how modern P2P technology tests the bounds of copyright law and content owners' available options to curtail full-fledged information anarchy on the Internet.

I. Freenet and the Rapid Evolution of P2P Technology

P2P technologies are spreading at an increasing rate, despite the persistent threat of legal action. Andrew Frank, Chief Technology Officer in Viant's Media and Entertainment practice, notes in The Copyright Crusade6 that the "ironic effect" of the content industry's vigorous protection of copyrights in court was to create a "Darwinian force ... under which only the most legally and technologically resilient [file trading systems] survive."7 P2P developers today focus on making copyright enforcement increasingly difficult, in addition to other technological goals.

Napster, Inc. believed it could escape secondary copyright liability by not directly transferring music files, but after losing in the Ninth Circuit in A&M Records v. Napster, Inc. ("Napster "), the service effectively shut down.8 In response, "FastTrack" based P2P technologies9 employ new distributed and encrypted "supernode"10 servers to index file names and facilitate searching.11 In addition to the technological robustness of distributed file indexing12 , it appears that the technology was also chosen to avoid the liability found in Napster . The content industry has wasted no time in pressing the issue of whether "FastTrack" companies can be liable under Napster , already filing suit against Grokster, Morpheus (StreamCast) and KaZaA.13

In contrast to the popular "FastTrack" technologies, many P2P developers are not focusing on "turning a buck," but rather on creating systems of file trading immune to both physical and legal attack. Perhaps the most advanced P2P technology in this regard is Freenet. Although the Freenet Project remains in the early stages of development, it has already created a secure P2P network that can withstand existing physical and legal attempts to shut down the network as a whole.

In July of 1999, Ian Clarke conceived of the idea for a secure, distributed file system during his studies of artificial intelligence and computer science at the University of Edinburgh, in Scotland.14 The idea quickly progressed into an open source, volunteer-supported effort, known as the Freenet Project. Freenet was publicly released in March of 2000 and has been under frenetic development ever since.15

A. Freenet's Technology

Freenet is not a "pure" P2P system (like Gnutella16), but is instead a distributed, caching file service.17 The Freenet program includes a client that retrieves files from other Freenet user's computers ("nodes") and a server that relays requests and files for other Freenet clients on the network. The Freenet network operates without the aid of any central technology.18 Each node of Freenet is autonomous and an "equal" on the system. Moreover, each node is aware only of nodes with which it can directly communicate. If a node disappears from the network, Freenet routes around this and the network continues unscathed.19

Unlike most P2P models, Freenet users who wish to share a file do not simply make a "shared" file available to other users and the network, like Napster or Gnutella do. Freenet users must "insert" the file into the Freenet.20 Users create a "key" for the Freenet file such as "freenet:the_constitution.txt," and insert the file into their node. The file is then stored on one or more local Freenet nodes. As Freenet users request the new file, neighboring nodes make additional copies of the file, distributing it across the Freenet.21 More popular files spread to more and more servers across the Internet. However, since Freenet servers have a limited storage capacity (determined by the node operator), less popular files get pushed out of servers and eventually drop off the Freenet altogether.22 Thus, Freenet does not rely on specific nodes to serve requested files, but instead acts as a large "cache," bringing more popular files closer to the users who want them.

In Napster , the recording industry was able to document the widespread infringement of copyrighted songs because all of the communications between the central Napster server and clients were sent unencrypted across the network.23 By contrast, Freenet encrypts all communications between nodes, so that third parties cannot monitor the content of file requests.24 Additionally, files are inserted in encrypted form. Only after a file successfully reaches a requesting Freenet client's hard drive is the file decrypted into an identifiable form.25

Freenet also takes pains to protect and guarantee the anonymity of its users. In contrast to Napster, there is no central database of users, or even a concept of "users." Each node only knows the Internet Protocol ("IP") address of its neighbors.26 When a file is retrieved or transmitted to a Freenet client, only the last node that contacts a monitoring node might be identified. However, there is no way of knowing whether that last node originated the file, or just passed on a "cached" copy from an earlier node. As files are requested and cached across the network, a node operator's server storage ("datastore") is used without the node operator's knowledge or control. A node operator cannot remove or determine what files are being served off of the node at any given point.27

B. Freenet's Organizational Structure

The Freenet Project aims to be robust organizationally, as well as technologically. Whereas Napster and others provide an obvious legal target of a central corporation to sue, Freenet is not the product of any specific company. The project currently consists of volunteer programmers, working on Freenet in their spare time.28 Moreover, Freenet is an open source project, meaning that anyone who wishes to take the existing source code of the project and continue development on her own may do so unimpeded. Thus, Clarke contends, "Freenet is an open, democratic system which cannot be controlled by any one person, not even its creators."29

C. Implications of Freenet

The Freenet Project developers focused on four primary goals for their technology: to make it highly survivable, private, secure and efficient.30 Freenet is designed to make it almost impossible to "shut down Freenet without shutting down the Internet."31 The redundant caching of information and distributed system provides a network with no dependencies on a central server, or any node in particular. Simulated tests have shown that the Freenet system "is surprisingly robust against quite large [network] failures," working efficiently even when 30 percent of the system nodes were randomly removed.32

Freenet protects the anonymity of users inserting files into the Freenet and other users requesting and retrieving those files (through encrypted transmissions). Moreover, the node that delivered a file is likely not the node that originated the file. Thus, Freenet makes all information of activities (sharing and receiving) within in its scope private from the many prying eyes of governments, copyright owners and other interested parties.33 Clarke characterizes the system as "a near-perfect anarchy."34

Freenet was originally conceived as a more efficient way to distribute information, later lending itself to the political goals of creating a haven for free speech.35 The technological nuances of Freenet that yield particularly troubling legal quagmires also contribute to the Freenet system's very effective distribution of data. By dynamically caching data, files are copied to relaying servers. Subsequent requests do not have to go quite so far to get the file, as it is now available from any of the relaying nodes, saving substantial amounts of bandwidth. Simulated tests and analysis of the current network suggests that the Freenet model will be able to handle an enormous level of traffic and activity.36

D. Status of Freenet Today

The Freenet Project has progressed rapidly from Clarke's original concept paper to a working software program and network. Downloads of the software have already topped 100,000, while it is estimated that a new 1,500 users download the software every day.37

Despite growing popularity, Clarke is quick to point out that Freenet is not the "next" Napster because "Napster is an application, but Freenet is a platform."38 Freenet is an architecture, which, like the Internet itself, can be used by programs for not only file sharing, but practically any form of communication. Developers are actively working on a variety of applications to take advantage of the security and distributed power of the Freenet network. The Freenet Message Board ("FMB") allows messages to be freely exchanged over the Freenet.39 Far more ambitious is the "Everything Over Freenet" Project, which aims to implement a variety of services, such as email, news, chat programs, and DNS40 , over the Freenet.41

Nonetheless, the file-swapping public and recording industry are most interested in how easily users can trade content files over the Freenet. In this regard, content owners can breathe easy. At the current time, Freenet is a limited tool for everyday users. Freenet is still quite difficult to install and the network is very slow.42 Moreover, Freenet currently lacks "free text" searches.43 A Freenet user has to know the exact file name "key" to retrieve a given file. Searching for the key "freenet:theconstitution.txt" will not match to "freenet:The Constitution.txt" or any other variant that is not the exact file key, character for character.

E. P2P and File-Sharing Trends

Riding this wave of popularity, P2P systems continue to develop and adapt to both technological and legal changes. New P2P systems are sprouting up all over the Internet.44 Moreover, broadband technologies, which allow greater download speed and use of these systems, are also expanding rapidly in the home and business environments. Freenet may not be the secure P2P system that ultimately gains a wide public acceptance, but it certainly serves as a conceptual model for future secure P2P technologies.

The large audience of file sharers are just as adaptive as the technologies they use, easily shifting to new P2P systems that better serve their needs. At the same time as the Ninth Circuit effectively shut down Napster, MusicCity offered its new FastTrack-based P2P client, Morpheus.45 Morpheus was an instant success and has now surpassed even Napster's load of simultaneous users.46 Andrew Frank observed that Morpheus "illustrates how, despite user interface woes and a hostile legal climate, a new peer-to-peer service can still double its usage in the span of a month and emerge from nothing to become a hit."47 The public has extremely low "switching costs," having no apparent problem abandoning their current P2P application for a new one, and is thus likely to migrate to the best P2P system, whether it respects copyrights or not.48

II. The Current Digital Copyright Regime

Copyright law has been the traditional sword the content industry wields against new technologies which are perceived as a threat.49 The Internet may be the darling new technology of the day, but the present P2P situation is far from the first battle that the content industry has waged.50 The tour of the modern confrontation between technology and copyright owners began in 1976, with Sony Corp. of Am. v. Universal City Studios, Inc. (hereinafter the "Betamax " case).51

A. The Betamax Case

Sony sold "Betamax " recording devices and tapes that allowed the public to record television programs.52 Two moguls of the television content industry, Universal City Studios and Walt Disney Productions, quickly brought suit against Sony for secondary copyright infringement.53 Universal and Disney contended that Sony should be liable for the copyright infringement of consumers who utilized the Beta system, which made the simple copying possible.54

The Supreme Court decided Betamax in 1984. In a 5-4 decision, the Court held that Sony was not liable for secondary copyright infringement because Beta was an "article of commerce," which was "capable of substantial noninfringing uses."55 The Court further reasoned that consumers who recorded television shows to play back at a later time ("time-shifting") were engaging in a legitimate "fair use", even if the copying was unauthorized by the copyright owners.56 Although a narrow decision, Betamax had far-reaching legal consequences. The doctrine of "substantial noninfringing use" has now become the standard by which to judge potentially infringing technologies and products.

B. Napster

In 1999, the popular success of Napster brought an almost immediate lawsuit against the file swapping service in A&M Records, Inc. v. Napster, Inc. (hereinafter "Napster ").57 The recording industry alleged that the vast majority of the MP3 files traded by the millions of Napster users were unauthorized copies.58 Napster responded with several theories under which the activity that Napster supported was defensible. Both the district court and Ninth Circuit sided with the recording industry.59

Napster argued that actions of its users in swapping files were protected fair use under the theory that Napster users were engaged in "space-shifting" their music from CD's to computers, devices, etc. in the same way that Betamax users "time-shifted" television programs.60 The Ninth Circuit concluded that the "shifting" defense was only available where a technology" did not also simultaneously involve distribution of the copyrighted material to the general public. Time or space-shifting copyrighted material was only protected when essentially done for personal use."61

Napster was first charged with contributory copyright liability, because it had knowledge of, and materially contributed to its users' copyright infringement.62 Napster argued that Betamax's "substantial noninfringing use" doctrine protected them from contributory copyright liability.63 The Ninth Circuit considered Napster's possible noninfringing "capabilities," but concluded that the doctrine was unavailable to Napster because they had notice and actual knowledge of infringement.64 The court also found that Napster materially contributed to copyright infringement by providing software and the central server.65 Thus, the court sustained an injunction for plaintiffs on their contributory liability claim.66

Napster was also charged with vicarious copyright liability, because Napster "has the right and ability to supervise the infringing activity and also has a direct financial interest in such activities."67 The Ninth Circuit found that Napster's business model depended on attracting a large user base to gain revenue. Thus, the pirated music available on Napster directly benefited the defendant.68 The court then examined Napster's ability to supervise its users and found that it sufficiently could.69 The court observed, "The ability to block infringers' access to a particular environment for any reason whatsoever is evidence of the right and ability to supervise."70 The court held that Napster's police power "must be exercised to its fullest extent. Turning a blind eye to detectable acts of infringement for the sake of profit gives rise to liability."71 Because Napster had not adequately supervised users and directly financially benefited from copyright infringement, the Ninth Circuit concluded that an injunction was appropriate on vicarious liability grounds as well.72

C. Grokster

After Napster was saddled with the Ninth Circuit's injunction and the once thriving file-swapping service lost its luster, the public found KaZaA73 , Morpheus74 and Grokster75 (hereinafter "FastTrack" technologies), based on the proprietary "FastTrack" technology.76 A FastTrack network relies on a collection of "supernodes" to provide listing and searching functions to users (like the old Napster central server).77 Unlike Napster, actual users run supernodes, and not a FastTrack company.78 The FastTrack companies claim that they have no knowledge of what goes on between the supernodes, as they do not directly control them and all supernode communications are encrypted.79

Following an increasingly standard approach to P2P technologies, the recording, music publishing and motion picture industries filed suit against the companies responsible for KaZaA, Morpheus, Grokster and FastTrack in 2001 (hereinafter the "Grokster " litigation).80 Although the Grokster lawsuit is still in preliminary phases at the time this article was written, the filings and motions of both sides provide a good indication of the eventual legal showdown. The parties have filed various motions and responses. As of December, 2002, both sides have argued summary judgment motions in federal court, hoping to prevail before the case actually goes to trial.81

The content industries claim that the Grokster defendants profit from users' copyright infringements like Napster, and are presently suing for both contributory82 and vicarious copyright liability.83 The plaintiffs allege that FastTrack defendants have actual knowledge of user infringement and that they actively encourage the public to trade unauthorized copyrighted works.84 The plaintiffs further contend that the defendants control users' copyright infringements because they reserve rights to unilaterally terminate user accounts.85 Finally, the plaintiffs argue that the defendants have derived financial benefit from the infringement of the content industries' copyrights.86

Defendant StreamCast's87 motion for partial summary judgment hints at the Grokster defendants' strongest arguments against secondary copyright liability under Napster : first, the defendants were running a network that was capable of "substantial non-infringing uses" protected under Betamax , and second, the defendants do not maintain a central database of available files like Napster.88

StreamCast initially noted that its Morpheus network has several non-infringing uses, leveraging the fact that Morpheus allows the exchange of any kind of document type (as opposed to only MP3's on Napster). One of Morpheus's non-infringing uses is Project Gutenberg's distribution of public domain "eBooks."89 Another current use of the Morpheus network is the availability of government documents.90 Third, copyright owners utilize Morpheus to distribute authorized and protected media.91 Finally, Morpheus allows the authorized exchange of demonstration, shareware92 and freeware computer software.93 Moreover, StreamCast concluded that if the Morpheus system could not utilize a Betamax defense ("staple article of commerce" defense), then liability would have to extend to most common internet technologies (like web servers, e-mail, FTP programs, etc.).94 In opposition, the content industry contended that these "legitimate" uses are insubstantial in the face of widespread piracy and that most often such files cannot be found on the defendants' networks.95

StreamCast also argued that because it does not operate a central file-index server, it could escape liability under Napster .96 Since FastTrack networks rely on user-controlled "supernodes" for file indexing, the Grokster defendants retain no control over the network or any supernodes.97 Moreover, StreamCast observed, "If [StreamCast] ceased to operate, or if its servers became inoperative . . . the searching, indexing, transferring, downloading [and other] functions . . . would continue unaffected."98 The content industry countered that the Grokster defendants retain control and knowledge of the content on their networks. Initially, the Morpheus program can recognize and determine similar files, which implies filters could be used to prohibit downloads of copyrighted works.99 Moreover, StreamCast regularly updates the software100 , sends promotional and other information101 , and monitors its bulletin boards102 from a central location. Finally, the language of the Morpheus end user license proves that "Morpheus is an ongoing, interactive service that [StreamCast] can terminate at any time."103

StreamCast contends that for a product to avoid secondary copyright liability under Betamax , it "need merely be capable of substantial noninfringing uses."104 StreamCast argues that "capability" was the crux of the Supreme Court's holding, and not the amount or degree of actual copyright infringement.105 Thus, StreamCast concluded it would be protected because "Morpheus software is capable of present and future substantial noninfringing uses."106 The content industry plaintiffs responded that these uses were not significant because the overwhelming majority of files traded on the FastTrack networks are illegal copies of copyrighted works.107 Moreover, since the Grokster defendants have actual knowledge of infringement and the FastTrack system was designed solely to circumvent copyright liability, the defense is unavailable.108

Turning to the Napster decision, StreamCast contended that the Ninth Circuit only held Napster, Inc. liable for its conduct, and not its architecture.109 Because StreamCast does not have knowledge of files indexed on supernodes and does not "'operate' the Morpheus user network," this architectural distinction precludes copyright liability under Napster .110 The content industry responded that Napster precludes any Betamax defense because of actual knowledge and control.111 The Central District of California denied StreamCast's motion and the case is set to go to full trial. Nonetheless, these arguments certainly provide a strong indication of the defining legal issues for the trial phase. Moreover, StreamCast's approach gives insight into the shifting legal landscape surrounding evolving file-sharing technologies. P2P projects, like Freenet, are mindful of the shaky legal ground on which they operate, and have adjusted to the Napster decision. The outcome of the Grokster litigation will likely further influence the design approaches of law-defying P2P technologies.

III. The Future of Copyright and Evolving P2P Technologies

Freenet makes an interesting study as a technological solution to the legal problems of censorship on the Internet. Ian Clarke's goal of destroying copyright law, as it conflicts with free speech, is an equally tantalizing legal subject, given that legal systems tend to be inhospitable towards deliberate attempts to circumvent the law.

Several technical aspects of Freenet have perplexing legal implications. First, any attempt to monitor activity on the Freenet will be limited and difficult. Freenet's encryption and communication methods make observation of a node or the network impracticable. As a result, monitoring and documenting wide-scale copyright infringement, like in the Napster case, is presently impossible with a system like Freenet.112 Second, Freenet's distributed "caching" system prevents data from being forcibly removed. If a court orders the takedown of material, a single node might be able to comply, but there is no way to filter or enforce the restriction on the network as a whole.113 The caching system also makes it virtually impossible to identify the source of the original document.114 The owner of a node serving up a file could be the original poster of the document, or the node could simply be one of thousands of nodes that unknowingly "cached" the document. Third, Freenet has no technical point of control. It was designed from its inception to be impossible to shut down, even by its own creators.115 Freenet nodes can survive on a hostile system, even when other nodes are shut down or turned maliciously against the network.116 Thus, legal or technological action against Freenet may prove fruitless against the system as a whole.

Additionally, it is worth considering that although copyright law may be effective at restraining P2P companies, it has done little or nothing to stop "direct" infringers - the millions of users trading files, with virtually no regard for copyright law.117 Moreover, these users move seamlessly from one P2P system to the next, as demonstrated in the rapid rise of the FastTrack system from Napster's ashes. Thus, regardless of the specific technology (Morpheus, Freenet, etc.) shut down next, the core dilemma remains, as the public's insatiable thirst for file trading is met by new and evolving technologies.

A. Enforcing Copyrights Against Individual Freenet Users

In Napster , although the recording industry did not pursue legal action against individuals, it had no difficulty discovering evidence of Napster users' "direct" copyright infringement by monitoring the Napster network. Unfortunately, even this is wrought with technological, legal and practical difficulties on a system like Freenet. Freenet's encrypted storage and communications, and lack of central communication point, prevent any party from monitoring general activity on the network. Furthermore, Freenet traffic looks very similar to other services, hindering the determination of which computers are even running the Freenet, let alone what files they are sharing.118

The Freenet system also complicates the individual user liability model under Napster . Clearly Freenet users who illegally download copyrighted works are guilty of infringing reproduction, but what of the nodes that serve the content? In Napster , individual users chose to share files they had stored on their computers, presumably infringing copyrights in distribution. By contrast, Freenet node operators never know what content is actually cached on their nodes, and have a plausible argument that they had no part in any actual infringement.119 Fred von Lohmann, an attorney with the Electronic Frontier Foundation, opines that Freenet's automated caching system, "minimize[s] the likelihood of direct infringement liability" for node operators, because it was designed for efficiency reasons.120 Moreover, the process by which a copyright owning plaintiff would actually prove individual infringement complicates the situation even more. Because Freenet replicates popular data, if a plaintiff requested and received a file from a neighbor node, it is very likely that the plaintiff's very request actually placed the file on the offending node!121 Thus, Freenet's caching system, where users and computers no longer control and own files, makes legal action against direct copyright infringers difficult.

In addition, even if individual liability were legally plausible, copyright owners could not go after the most wanton infringers. Traditionally, a P2P system (like Napster or Gnutella) is driven by small groups of users (the "big dogs"), who provide most of the files. While it is infeasible to sue everyone on a P2P network, going after the major suppliers of information is a somewhat effective legal strategy.122 However, this tactic is completely ineffective for Freenet, because node operators do not control or know what is on their node. Someone else most likely inserted the original infringing file, and eventually got "cached" at some given user's node. On Freenet, node operators cannot be reliably identified and distinguished as saintly, copyright-obeying users versus willful, copyright pirates and distributors.

B. Enforcing Copyrights Against Organizations Promoting Freenet

Copyright owners usually forego legal action against individual infringers, concentrating instead on identifiable companies releasing the technologies. However, Freenet is difficult to target as an organization. Freenet itself is not "owned" by any one entity. The Freenet Project does not operate for profit, and relies almost entirely on the volunteer efforts of developers. Freenet is an open source project, allowing anyone to take the source code and continue independent development.123 Andrew Frank sums up the difficulty in initiating legal actions against such projects:

Going after the creators of such services does little good: first, these systems are generally being developed using Open Source collaboration, so hundreds of anonymous developers worldwide may contribute to the final product; second, even if these alleged perpetrators could all be identified and prosecuted, they would not be able to stop the service once it was unleashed, as it would be running on the individual computers of its users.124

Professor Jessica Litman, in Digital Copyright , echoes this sentiment, noting, "As a comprehensive strategy, litigation works best against commercial actors."125 Litman concludes that a litigation strategy will prove increasingly difficult for programs that offer no clear intermediary to sue, like The Freenet Project.126

Nonetheless, it is possible (some would say even likely) that Freenet developers could eventually become a legal target. However, traditional copyright law does not lend itself well to Freenet, even in a Post-Napster world. Initially, proving end users' direct infringement (required for secondary copyright liability) was relatively easy in Napster and the ongoing Grokster litigation. By contrast, the previous section shows why proving copyright infringement by individual Freenet users is impracticable and complicated. Nevertheless, von Lohmann concludes that although a P2P system is protected by "plausibly deny[ing]" user infringement, the promotion, endorsement or aid in the infringing use of a P2P product may nonetheless potentially garner secondary copyright liability.127

Proving the remaining elements of secondary copyright infringement against Freenet developers would be equally difficult. Vicarious liability does not really apply to Freenet. Vicarious copyright liability requires proof of a defendant's right and ability to control the copyright infringement of others and the derivation of a financial benefit from that infringement.128 In Napster , the Ninth Circuit found that operating the central file-indexing server and running user accounts left Napster in control of its system.129 In the Grokster case, the content industry alleges the defendants control their FastTrack systems through user accounts and rights to terminate users.130 Freenet, by contrast, has no concept of outwardly identifiable "users" or any form of centralized control. Moreover, Freenet derives no direct financial benefit from the project, and its current developers obtain no obvious present or future financial gain from Freenet.

Claims for contributory liability might fare a little better, but would nonetheless prove difficult against Freenet. Contributory liability requires that the defendant had knowledge of and materially contributed to direct copyright infringement.131 Knowledge can be demonstrated "by showing either that the contributory infringer actually knew about the infringing activity, or that he reasonably should have known given all the facts and circumstances (constructive knowledge)."132 In Napster , actual knowledge was shown by company e-mails and monitored lists of thousands of infringing files.133 Knowledge of Freenet developers would be difficult to prove based on evidence of monitoring files (like in the Napster case). However, if infringing files were found to be prominently listed on the Freenet, direct or constructive knowledge would be a feasible claim, especially given that infringing content largely comprises the current network.134 Additionally, Ian Clarke and other Freenet developers have publicly acknowledged the potential (and for some, the goal) of Freenet subverting copyright controls on file-swapping.135 Proving material contribution to direct user infringement would likely be a straightforward claim against Freenet developers who offer the program to the public.136 Thus, Freenet developers could conceivably face some form of contributory liability, although even such a finding still could not stop the Freenet network, or someone else from resuming their development roll.

The "substantial noninfringing use" doctrine (the "Betamax defense")137 could be argued by any Freenet group claiming fair use protection for the "legitimate uses" of Freenet. Initially, Freenet is championed as a protector of free speech. Andrew Frank, writing for the content industry, even admits that "[l]egally, however, because their content is encrypted, [distributed, encrypted P2P systems] have a substantial non-infringing objective: to protect freedom of speech from censorship."138 Although the copyright - free speech issue is a matter of ongoing debate,139 it is conceivable that the core goal of the Freenet Project could be found to be a "substantial noninfringing use." Second, in the Grokster litigation, StreamCast has already indicated it believes that Morpheus has several substantial non-infringing uses (distributing public domain content, authorized software and media content, etc).140 Any of these possible non-infringing distributions are possible on the Freenet as well. Third, Freenet is not necessarily an application that allows infringement, but a full "platform" in its own right.141 An application, such as Freeweb (which allows a user to view websites over the Freenet) or Frost (which allows users to trade files over the Freenet) is required to actually perform any file-sharing.142 While groups that develop these specific tools could face liability, the core Freenet network is likely legally immune because it does not actually swap files.143 Fred von Lohmann concludes that when P2P technologies disaggregate functions in this fashion, they stand a much greater likelihood being of legally immune.144 Fourth, the design choices of dynamically caching documents can be touted for their technological desirability, despite their hindrance to legal enforcement of copyrights. Caching and file distribution techniques save bandwidth, bring desired information more quickly to those who want it,145 and provide free storage space for information.146 Encryption of information guarantees data integrity against those who might attack communications. Additionally, the decentralized P2P system may prove more efficient at searching for information.147

Thus, the wide array of technological uses, free speech goals and potential for dissemination of "fair use," authorized, or public domain content leave Freenet in a better legal position than either Napster or the Grokster defendants. Nonetheless, the Napster decision (if followed by other circuits) has seriously limited the scope of the "Betamax defense" in the P2P context, and the mere copyright-subverting goal of Freenet could yield liability.148 However, even if a court refuses to find "substantial non-infringing uses" of the Freenet system, the content industry will inevitably have to face the prospect of future P2P developers adopting these legally troubling designs as a standard, efficient approach to their new technologies.

Notwithstanding traditional secondary copyright liability, a group connected to Freenet could possibly try to seek refuge under section 512 of the Digital Millennium Copyright Act ("DMCA"), but would not likely be successful.149 Section 512, the "safe-harbor" provision, provides "online service providers" with four bases of protection, covering: "[t]ransitory communications[,]" "[s]ystem caching[,]" "[s]torage of information on systems or networks at direction of users[,]" and "[i]nformation location tools."150 However, von Lohmann notes that because "Congress did not anticipate peer-to-peer file sharing when it enacted safe harbors, many P2P products may not fit within the four enumerated functions." Von Lohmann posits that the transitory communications exception151 would not apply unless the network traffic actually passes through the P2P's private network.152 Napster failed under this standard, and Freenet clearly would as well, given that Freenet has no "private" network of its own.153 Although the "system caching" exception154 seems to encompass Freenet, this and the remaining two exceptions all require some form of blocking access to or taking down information upon notification of copyright infringement.155 The Freenet system is thus problematic under DMCA safeguards because it deliberately prevents the forcible takedown of information.

Additionally, any group of Freenet developers sued by a plaintiff would likely have difficulty proving they are an "online service provider" ("OSP").156 To qualify for any of the safe harbor protections, an OSP must terminate users who infringe copyright laws and not interfere with "standard technical measures" that protect copyrights.157 Freenet does not employ a "user" system; the very design of Freenet as a censorship resistant technology could be seen as an interference with "standard technical measures."158 Also, von Lohmann notes that an OSP "must not have known about the infringement, or been aware of facts from which such activity was apparent (i.e., if you take a 'head in the sand' approach, you lose safe harbor)."159 The developers of Freenet would likely be found to have at least constructive knowledge of copyright infringement.160 Alternately, if the content industry attempted to sue internet service providers ("ISP's") (such as America Online, etc.), for merely allowing users to run Freenet nodes, it is likely that such ISP's would qualify as OSP's and fall within one of the safe harbor categories.161 However, for any group directly connected to Freenet, and merely aiding a Freenet project, the DMCA safe harbor protections would likely be unavailable.

C. Prospective Legal Approaches to Evolving P2P Technologies

The P2P legal landscape has been continually changing since Napster . As the content industry moves forward in its litigation with the FastTrack P2P companies, the FastTrack system may prevail where Napster failed because its service is more decentralized and encrypted than Napster. Furthermore, if the content industry does succeed in the Grokster litigation, P2P developers will move on, trying to create a system that cannot be stopped by copyright law.

Some observers have suggested that the DMCA could be amended to require ISP's or individual users to only allow P2P technologies over their networks that respected copyrights.162 Freenet developers have even pondered Freenet specific legislation or litigation that would impose burdens on individual Freenet nodes.163 However, such approaches would likely run afoul of the current "no mandate" technology provision of the DMCA164 , and restrain technological innovation too greatly to be palatable options.165 Moreover, Freenet (or other P2P) communications may eventually become so well-masked and undetectable that legal action would be practically moot.166

In the meantime, the content industry has not been waiting idly for the day when the existing copyright framework falters under technological advances. In addition to attempting to thwart existing P2P networks,167 the content industry has begun to push the government for stronger copyright controls on the user's end.168 Damien Cave observes the broad scope of the content industry's vision:

Hollywood is on the march. Adding copy protection to CDs is just one tactic in a comprehensive onslaught. Media behemoths like Disney, Sony and AOL Time Warner are seeking full control of all methods of entertainment distribution; if their vision is realized, digital television sets, hard drives, personal video-recorders and wireless devices will all have some form of copy protection. In the most dire incarnation of the digital entertainment future, consumers of music and movies won't be able to make any copies at all without explicit permission...169

The content industry is seeking the help of the government to force users away from "insecure" medium to copyright-owner friendly technologies.170 This approach, also called "digital rights management" ("DRM"), aims to simply make it impossible for any individual to play unauthorized content on their computer or device.171 The Consumer Broadband and Digital Television Promotion Act ("CBDTPA"), proposed by Senator Ernest Hollings at the behest of the content industry, would simply bypass the P2P threat by requiring all computers and digital devices to ensure that only authorized copyrighted content could be accessed.172 Although the CBDTPA has not made measurable progress in the Senate at the time this article was written, the content industry is steadily moving along with a mixed technology-legislation DRM approach. If successful, all existing P2P systems would cease to be a copyright threat if their unauthorized content could not be used on the end user's computer. Not without its critics, DRM has been harshly criticized for putting far too much power over information in the hands of the content industry.173 Additionally, many observers note that the DRM approach may alienate too many consumers, as well as ultimately prove technologically impossible.174

D. Prospective Non-Legal Approaches to Evolving P2P Technologies

Should the current regime of copyright law fail to stop Freenet, content owners or the legal system could attempt to technologically attack or poison the system.175 Nevertheless, this threat has been assumed from the start, and the Freenet Project maintains that it can only be shut down when the Internet is turned off.176 Architectural changes in the Internet could make this more feasible, however. Lawrence Lessig contends that the Internet as a whole is moving away from being a "stupid" network that is difficult to control to a "smart" network that will be more readily manipulated by powerful corporate and government actors.177 Although a remote threat as pertains to Freenet now, things change very quickly on the Internet. Such architectural change could eventually control the Freenet, which currently relies on the nondiscriminatory flow of data in the current "stupid" network design to hide and secure the Freenet network.

If both the technological and legal arenas fail to stop P2P systems like Freenet, the content industry may have to reconsider its approach towards P2P technologies. Professor Litman argues that content industry must take steps beyond their current legal and technological strategies:

If forty million people refuse to obey a law, then what the law says doesn't matter. It may be that people flout it because they're natural lawbreakers, or it may be ... that they don't comply because it doesn't make sense to them. Whatever the reason, the law is not going to work well in the real world.178

Professor Litman concludes that this is simply the age-old tradition of copyright owners resisting all new forms of technology they cannot control.179 Talal Shamoon, executive vice president of InterTrust, a company pioneering copy-protection strategies for the content industry, acknowledges that the P2P controversy illustrates an "ugly transition period" in the digital revolution.180 Nonetheless, Shamoon believes that regardless of a particular solution, the content industry will adapt to the specter of P2P technologies and other advances.181 As it turns out, Ian Clarke would agree with Shamoon's assessment: "Artists and publishers all adapted to those new technologies and learned how to use them and profit from them; they will adapt to Freenet as well."182

Conclusion

More legal questions and conundrums are raised with a technology like Freenet than can be currently answered. Presently, Freenet is still an enthusiast's toy and not the "next" or current Napster, Morpheus, etc. However, even if Freenet never gains a massive user base, the law-defying encryption and distributed caching techniques of the project will likely end up in the next generation of P2P services. The struggles over changes in the Internet, seen through the eyes of emerging technologies, demonstrate that the confrontations between copyright owners and free information advocates will only continue to escalate. This escalation will be inextricably tangled in both legal and technological complexity, as neither the law nor technology appears capable of solving these dilemmas alone.

As Andrew Frank poignantly observes, P2P technologies are evolving in a "Darwinian" fashion, proving more resistant to technological and legal control with each iteration.183 The content industry stopped Napster. The industry may stop the FastTrack companies. It may even stop Freenet. Eventually, however, a new system, borne of the lessons of these pioneering technologies, will likely arrive that cannot be addressed within the current practical confines of copyright law. When that day comes, the content industry will perhaps have to consider (if has not already done so) how it will "evolve" in the ever-changing digital landscape.

Footnotes