1. Introduction
“Piracy is a business model. It exists to serve a need in
the market – consumers who want TV content on demand. And
piracy competes for consumers the same way we do: through quality,
price, and availability.”1
Consumer
options for consuming creative content in digital form, such as
music, movies, books, and television shows, have increased
significantly with the development of the Internet. Entrants into
the distribution stage of production have developed new business
models to deliver digital content in general and copyrighted digital
content in particular. The objective of this paper is to analyse the
development of competition in the delivery of digital content to
consumers. In particular, the focus is on new technologies that
facilitate online dissemination of digital content to consumers
through the use of peer-to-peer (P2P) networks and video hosting
sites that have proliferated over the Internet since the late 1990s.
The role of copyright as a potential competitive weapon by incumbent
disseminators is joined in the analysis. P2P file sharing networks
and video sharing web sites are viewed as entrants into the market
for the dissemination of digital content to consumers. The incumbent
technologies for distributing content have reacted aggressively to
this new source of competition and have pursued legal, economic, and
moral strategies to combat the use of authorised and unauthorised
content by the distribution entrants. Perhaps the most important
point to keep in mind is that online distribution, both authorised
and unauthorised, is here to stay.
Section
II presents background information depicting a stylised view of the
value chain/production process and models the consumer’s
choice problem by identifying characteristics of online distribution
platforms that factor into the consumer’s decision making
process when the consumer accesses digital content. Section III
describes the varying technological designs that entrants utilised
to distribute digital content to consumers. Section IV analyses
legal strategies that copyright owners have pursued to address the
threat of competition and poses four questions that software
developers of new online distribution platforms need answered
clearly and with predictability to spur technological advancement.
The answers to the questions posed require an extended discussion of
the legal and economic issues surrounding the entrants’
technology and business models. Section V offers a conclusion.
2. Background and Consumer Choice
Problem
Diagram
1 is an overview of the stages in the production and delivery of
content. The production process/value chain in general includes the
creation of content, negotiations between the content creator and
the publisher of the content, the distribution of content for access
to consumers, and the consumption of content by consumers.
DIAGRAM
1. Stages of Production/Value Chain
Content
creators include artists who write and record songs, write
television shows, author books, and write movies. Historically,
content creators aspired to earn a living from the work, although
there is an increasing body of content that is now make freely
available to the public. Content is then published by music
companies, book publishers, movie studios or television production
companies that put the content in form consumers will find
interesting. Frequently, the ownership of the copyright ends up in
the possession of a firm who is also integrated into the
distribution stage of production. The distribution stage provides
the delivery service of content to consumers. It should be noted
that access to copyrighted and non-copyrighted content is an
essential input into the distribution stage of production.
Firms in
the distribution stage differentiate their service over a set of
characteristics/attributes that they hope is important to consumers
when they decide how to gain access to their preferred content.
Bakker identifies eight different aspects upon which paid download
services (such as Apple’s iTunes music store) and peer-to-peer
file sharing services (such as LimeWire and Kazaa) compete to
attract consumers.2
Sag develops a benefit-cost model comparing the net benefits from
unauthorised downloading to the net benefits from legally acquired
music.3
In particular, Sag categorises benefits in terms of functional
value, normative value, and law-abidingness value and costs in terms
of monetary cost, search costs, expected costs from computer
viruses, and expected cost of sanctions.
In
general, alternative online delivery services for content, such as
licensed services, unlicensed file sharing services, and video
sharing services, offer consumers different bundles of
characteristics from which to choose. Building on the Bakker and Sag
models, the following set of characteristics are likely to be
considered by a consumer when choosing between alternative
distribution services:
Price
Selection of content
Quality
Speed of access
Security
Legal liability
Ease of use/portability
Search cost
Unbundled access to content
Sense of community
Extra features
It should
be noted that each characteristic may not be relevant to a consumer
for all forms (e.g., music versus television show) of content.
Bakker
was interested in identifying the aspects/characteristics that were
of most importance when a consumer chooses between paid legal
download services and file- sharing services. His research revealed
that the online distribution market is composed of two distinct
types of consumers, file sharers and paid downloaders. File sharing
services are attractive to a younger, more computer savvy audience
that displays both patience and persistence, whereas the paid
download service attracts a more mature consumer. Thus, online
distributors need to understand this market segmentation indicating
two different types of consumers and develop appropriate business
strategies to target each segment. The focus of Sag’s model
was to analyse the impact on consumer behavior from a litigation
strategy that targeted marginal file sharers.
Felten
sketched a theoretical model of consumer behavior in which he
further disaggregated the market segment consisting of file
sharers.4
Specifically, he decomposed file sharers into two types, free riders
(young people with few moral qualms about file sharing) and samplers
(older, more risk-averse, people who were morally conflicted about
file sharing). This decomposition enabled Felten to explain
disparate findings in empirical studies examining the impact of file
sharing on music sales.
None of
these models factors into the consumer decision making process the
entry of video sharing web sites, such as YouTube and MySpace, which
concentrate their efforts on building a sense of community among
users of the service. These sites provide access to copyrighted
(often unauthorised) content and, more importantly, user-generated
content. The characteristics/attributes identified above provide a
conceptual framework for analysing the competition between the
content providers’ sanctioned online providers of content and
the new entrants in the form of peer-to-peer file sharing networks
and video sharing web sites.
3. Architecture of Entrants’
Networks into Online Distribution
A
distinguishing characteristic of a P2P file sharing network is that
content is not stored at a server within the network but rather on
peer computers at the edges of the network. A user wishing to
utilise a P2P file sharing network needs a software application
program downloaded from a P2P software provider’s web site
that will enable the user to locate others users on the network, an
ability to locate content available from the network edges, and a
communications protocol to exchange files.
The
emergence of Napster in May 1999 introduced the world to a hybrid
version of a peer-to-peer file sharing network for distributing
digital content, in Napster’s case music MP3 files. Napster’s
network facilitated the sharing of (mostly unauthorised) music files
in unprecedented magnitudes. Napster’s architecture contained
a central server that indexed the files that were available on peer
computers from the network. The most noteworthy P2P innovation of
the network was decentralisation in storage of shared files.5
Individual computers at the edges of the network hosted the sought
after content. So when a user downloaded a file, the copy of that
file was uploaded from a peer in the network, not a central server.
Thus, Napster facilitated the exchange of music files by providing a
central index of the content and arranging for the connection
between the downloading and the uploading peer computers. This
architecture, containing a centralised index, meant legally that
Napster exercised “control” over the network and when
copyright holders notified Napster that copyrighted content they
owned was available on the network without their permission, they
than possessed “actual knowledge” of infringing
activity. Control and actual knowledge led the court to conclude
that Napster was liable for indirect copyright infringement because
its service facilitated direct infringement by its users. The
“original” Napster eventually shut down operations in
2001 after attempting to install filtering mechanisms.
Software
developers of peer-to-peer file sharing technology learned key
design lessons from the Napster decision. Developers began to
disaggregate the functions necessary (such as search for files and
sources of downloaded files) for a P2P network. P2P file sharing
networks were introduced that did not contain a central index of
available content and developers relinquished control over the
behavior of the peers in the network once a peer downloaded the
client software, hoping to have the courts view the software as a
product (like a VCR) and not a service capable of an ongoing
relationship. Companies such as Kazaa, Grokster, and LimeWire
emerged as successors to Napster in terms of facilitating the P2P
sharing of content, mostly presumptively copyrighted. Diagram 2
shows alternate network designs.
DIAGRAM 2. 6
Pure Peer Network (early Gnutella)
Hierarchical Peer Network (Kazaa, Grokster, Lime Wire)
Centrally Coordinated Peer Network (Napster)
Client Server Network (YouTube, MyMP3.com)
An
example of a pure P2P file sharing network was the original design
of Gnutella (released March 2000) in which the search function and
content storage were totally decentralised, meaning that each
function was conducted at the individual peer level. This design
suffered from several technical weaknesses that have diminished its
role as a competitive distribution platform.
The
client server network, utilised by emerging video sharing web sites
such as YouTube and Grouper and licensed distributors such as
iTunes, is at the opposite end of the spectrum in terms of control
for peer interaction only occurs with a central server and the
server hosts the content and provides search functionality. Thus,
there are no P2P aspects to this design.7
The centralised control of all functions is an operational strength
of the design. Somewhere between the two polar designs is the
original Napster’s centrally coordinated peer design with
centralised search functionality but peer storage of content. Also,
in the middle of the spectrum is Kazaa and Grokster’s file
sharing networks that feature a hierarchical peer design. The search
functionary is carried out, not centrally, but rather at peer
supernode computers and files are hosted and exchanged through
supernodes among peers. There are benefits and costs to each design
but the focus of the paper is how legal decisions influence the
design of network architectures that are emerging as competitors in
the online distribution market. Wu characterises the behavior of
developers who design application software with this overriding
architectural objective as “…designing code to avoid
copyright infringement.”8
In 2001,
copyright owners sued the P2P file sharing companies Grokster and
StreamCast Networks (makers of the Morpheus software) that utilise a
hierarchical peer design and won a significant legal decision in the
June 2005 Supreme Court’s Grokster decision. The
Supreme Court did not evaluate in its main opinion the merits of the
hierarchical peer design but instead imported from patent law the
theory of active inducement, applied it to the facts of the case,
and found reason to remand the case back to district court. The
decision means that for future copyright cases of this nature, first
the court will examine the intent behind the development of a new
technology before it proceeds to determine whether a dual use
technology that facilitates both infringing and non-infringing uses
is legal. The Grokster decision hinged on the “bad”
behavior of the companies rather than on the architectural design of
the software they distributed.
In
September 2006, a set of copyright owners sued in the southern
district court of New York the P2P file sharing company LimeWire
that offers client software that utilises the revised Gnutella
network hierarchical design with a decentralised search index. Thus,
the court may get another chance to analyse LimeWire’s file
sharing technology design using the Sony decision of 1984 and
its “capable of substantial non-infringing use” rule.
In
response to the Napster and Grokster legal decisions,
P2P file sharing software developers have continued to adapt P2P
code. Choi examines how P2P file sharing developers continue to
decentralise the functions of P2P networks.9
Specifically, Choi characterises the tactics of file sharing
software developers to disaggregate the search and delivery
functionalities to insulate the networks from legal challenge as a
“guerrilla movement” against copyright owners. For
example, BitTorrent file sharing software provides an efficient way
(swarming) to share (especially large such as video) files among
peer computers and discourages free-riders through a tit-for-tat
element of the protocol but at a time and patience cost for users.10
A downloader must first find a web site that lists “torrent”
files. A “torrent” file contains metadata that contains
information about the location of a computer that hosts the
particular file and the location of a “tracker” server
that is concurrently coordinating exchanges of the file. The
downloader then clicks on the requested file to enter the exchange
process. Search engines such as Isohunt, Torrentspy, Torrentbox, and
Pirate Bay provide directions to files that are shared across the
network.
The
disaggregation of the search and file transfer functions make the
process of file sharing for users more complex and costly in terms
of time but to date less susceptible to legal challenge. Although
copyright owners have shut down some of these tracking/index sites,
Choi argues persuasively that, “While trackers can be shut
down and removed from the Internet, this process is about as tedious
as shutting down individual direct infringers.”11
That is, the comparative cost advantage of pursuing intermediaries
as indirect infringers compared to direct infringers diminishes as
the networks become increasingly decentralised.12
Thus, the cost effectiveness of continuing lawsuits alleging
indirect liability is in doubt. On the other hand, the more involved
search process increases the relative cost of utilising an
unauthorised P2P file sharing program to an end user. This may
persuade consumers to turn to legal download services for content.
Video
sharing web sites such as YouTube use a centralised server client
network and have emerged as significant competitors in the
distribution stage of production of content, copyrighted and
user-created. Users upload content for aggregation on the web site
and the site provides a search function for users. YouTube is
different from existing distribution channels for it is designed to
accommodate a more participatory role by end users in modifying and
enhancing existing content as well as creating new content. End
users are not expected to be mere passive consumers of static
content. Sites such as YouTube hope to utilise Section 512,
contained in the Digital Millennium Copyright Act of 1998, for
insulation from legal liability.13
Section 512 seemingly provides a safe harbor for web sites that
store materials on behalf of users as long as they follow a set of
legal and technical requirements for removing infringing material
when notified by the copyright owner. Nevertheless, in many cases,
these sites end up hosting infringing material. YouTube has actively
sought licensing deals with copyright owners to reduce the
likelihood of lawsuits even though it presumptively qualifies for
the safe harbor.
Diagram
3 identifies representative competitors for three alternatives
network designs that provide online distribution of content:
Diagram
3. Representative Competitors
|
Licensed/Authorized Online Distributors
|
Unlicensed/Unauthorized P2P File Sharing Distributors
|
Video Sharing Websites
|
|
iTunes
eMusic
Yahoo Music
Rhapsody
Zune
MovieLink
CinemaNow
Vongo
Guba
(new) Napster
iMesh
Peer Impact
Shared Media
Mashboxx
|
Bit Torrent
LimeWire (currently sued and countersued)
Grokster (shut down)
eDonkey (shut down)
WinMX (shut down)
i2hub (shut down)
Napster (shut down)
Aimster (shut down)
|
YouTube
MySpace
MSN
AOL
Yahoo
blip.tv
Veoh
Joost
Grouper
Bolt
TVV
Revver
|
Sources: (i) Center for Democracy and Technology,
Post-Grokster Secondary Liability Developments,
January 2007.
(ii) Andrew Currah, “Hollywood versus the Internet: the media
and entertainment industries in a digital and networked economy”
(2006) 6 Journal of Economic Geography, 449.
(iii) Various issues of the Wall Street Journal.
The arrow
next to BitTorrent indicates that the company who developed the
open-source communications protocol that is widely used by P2P
clients to share files is now soliciting licensed copyright deals to
transform itself into an authorised online distribution medium.
The
diagram indicates an increasingly competitive environment in the
distribution of online content. Generally, as competition increases,
consumers benefit from lower prices, increased output, and a more
rapid pace of technological change. To stem the tide of entry into
distribution, the extant content distributors are pursuing an
aggressive legal strategy with multiple tactics. On one hand, in the
view of copyright owners, the goal of such a legal strategy is to
prevent unlicensed competitors from utilising intellectual property
without permission or authorisation. On the other hand, such a
strategy tends to increase the uncertainty on the part of technology
developers as to the legality of innovative methods of online
distribution and, thus, may tend to reduce innovation in the
market.14
In order to examine potential sources of uncertainty, an analysis
of the legal and economic environments facing technological
innovators is necessary.
4. Legal and Economic Environment
4.1 Indirect Liability Strategies
Online
entrants offer services that are utilised by customers who engage in
both non-infringing and infringing activities. Direct copyright
infringement may occur when a customer reproduces, distributes,
publicly displays, creates derivative work, or publicly performs
copyrighted content without the authorisation or permission of the
copyright owner. Indirect copyright infringement may occur when the
developer of a product/service/device utilised by a customer
facilitates the unauthorised direct infringement of copyrighted
content. To build a successful business, researchers and developers
of new dual use technologies rely on established legal rules for
assurance that their product is immune from indirect liability. That
is, in order to provide a legal environment conducive to innovation,
it must be clear under what circumstances a developer is responsible
for the direct infringement acts of its customers. The following
four questions provide a framework to understand the settled ground
rules for competition as well as the unresolved issues for which
uncertainty still exists.
QUESTION
1: Is the product/service capable of substantial non-infringing
uses?
Suppose a
developer creates a product, such as a photocopier, MP3 player,
computer, or digital video recorder that can be used by its
customers for both non-infringing and infringing uses. One key legal
issue is whether or not the developer is indirectly liable for the
acts of direct infringement by its customers. Indirect liability, in
general, asks whether the developer created a product to
intentionally encourage or induce direct infringement and profited
from that direct infringement. The Supreme Court addressed this
issue in 1984. The Sony rule, created as a result of a
lawsuit between Sony and Universal City Studios involving the analog
videocassette recorder (the Sony Betamax), concluded that “the
sale of copying equipment, like the sale of other articles of
commerce, does not constitute contributory infringement if the
product is widely used for legitimate, unobjectionable purposes.
Indeed, it need merely be capable of substantial non-infringing
uses.”15
Thus, a developer of a technology that a customer may use to
infringe directly on copyright must be confident that the product is
also capable of substantial non-infringing uses. If this is
demonstrated, the product is free from liability even though the
developer knows or has reason to know that the product can be used
for infringing purposes. If it is not capable of substantial
non-infringing uses such as in the case of cable descramblers, it is
likely that the developer of the product will be subject to
liability for indirect infringement and thus in the long run its
business model is not likely to be economically viable. The
extension of this rule from an analog device such as a videocassette
recorder or a copying machine to digital services offered by Napster
or peer-to-peer software providers such as Grokster, Lime Wire, or
BitTorrent, and perhaps to video sharing sites such as YouTube, has
proven both interesting and controversial.
There
still exists uncertainty with how words in the Sony rule
should be or will be interpreted. Is there a need to quantify the
meaning of “substantial” and should or will a future
court set a benchmark for the ratio of non-infringing uses to
infringing uses? It appears based on the opinions in the Sony
and Grokster16
cases that a product must have at a minimum ten per cent of its uses
non-infringing to be safe from indirect liability. The plain meaning
of the word “capable” suggests that one should count
current uses and potential future uses, although one circuit court
argued that one should only include probable uses.17
Which party in litigation bears the burden to demonstrate whether
the product is or is not capable of substantial non-infringing uses?
Finally, and an issue to be explored in more detail below, should
the developer of the technology be responsible to have considered
alternative cost effective designs of the product that would have ex
ante reduced or eliminated infringing uses?
QUESTION
2: Does the design of the service enable the developer to have
actual knowledge of infringement and control to eliminate or reduce
economically the infringing uses of the product/service?
Peer-to-peer
software developers learned several key lessons from the litigation
over Napster’s design. Napster’s architecture
decentralised the storage of files but maintained a site and
facilities that included a central server that indexed music files
that were available to be shared by its customers. This design
subjected Napster to charges of facilitating direct infringement by
its customers. Copyright owners successfully sued Napster for
indirect infringement and the decision rested on the ability of
Napster to eliminate the infringing uses of its service when it
received actual knowledge of direct infringement activity by its
customers. The court concluded that Napster could and should
redesign its service to block the exchange of any infringing music
files while continuing to allow permissible uses.18
The centrally coordinated peer network employed by Napster gave it
control over the acts of direct infringement by its customers and
copyright owners gave Napster notice of the infringing acts of its
customers at a time when Napster could do something about it. Thus,
peer-to-peer developers learned to create software that could not
easily be redesigned to give control over customers or actual
knowledge of infringing activity. In short, the lesson learned was
to design code to relinquish control over peer-to-peer software and
thus, how consumers used the software.
This led
to the introduction of the hierarchical peer design. Based on the
rules learned in response to the first two questions, developers
were led to believe that, if they developed a service that was
capable of substantial non-infringing uses, utilised a design that
decentralised the search function, stored files at peer computers,
and had no ability to reduce or eliminate direct infringing acts,
then the service would be immune from indirect liability. Moreover,
if the developer’s company ceased business, customers could
still use the software to share files. But, in a surprising decision
to the legal and technology communities in 2005, the Supreme Court
created a new theory of indirect copyright liability.
QUESTION
3: Did the developer of the service engage in behavior to actively
induce copyright infringement?
The
significance of the Grokster decision is that the behavior of
the developer may be subject to indirect liability under an active
inducement theory. Specifically, a developer’s actions and
motives are scrutinised for signs of “clear expression or
other affirmative steps taken to foster infringement.” If the
developer’s firm is found liable for inducement infringement,
than a service’s design and capability for substantial
non-infringing uses are irrelevant. This means that a developer
cannot design its dual use service to eliminate control and actual
knowledge as Grokster did and at the same time take active steps to
solicit customers based on the potential for the service to infringe
copyright. The court identified three possible elements of intent to
infringe:
advertise the ability of the service to infringe on copyright,
fail to proactively filter out infringing uses,
rely on a business plan that is directly linked to the volume of
infringing activity.
According
to the Court, element ii or iii alone is not sufficient to establish
intent to infringe.
There are
several outstanding issues concerning interpretation of the rule on
active inducement. First, in the case of Grokster, the Court found
evidence of all three elements of inducement and concluded that
there was “bad behavior” by the defendants. It is not
clear whether all three elements must be present or perhaps only two
of them or element i alone for “bad behavior” and thus
intent to infringe liability. It is unknown the range of behaviors
(e.g., anti-spoofing features, use of encryption, private viewing
groups) that could be considered evidence of inducement.19
Hopefully, according to one prominent legal scholar, courts will
apply high standards to demonstrate inducement liability.20
Second, elements ii and iii are not necessarily independent. If a
developer planned on building a business model based predominantly
on facilitating infringement, the developer would not likely filter
the service. Thus, the conclusion that each element ii and iii alone
is insufficient for inferring intent to infringe is based on an
incomplete analysis. Third, there is no quantification standard of
what it means to rely on infringement as a major revenue source in a
business model. Most services will experience an increase in revenue
from increased use of the service regardless of whether it is from
non-infringing use or infringing use. Fourth, the Grokster
rule emphasises affirmative steps suggesting that proactive
filtering efforts are unnecessary for immunisation. However,
copyright owners may try to expand the scope of the rule to infer
intent from design by arguing that a failure to filter proactively
is an indication of intent to infringe. Each of these unresolved
issues creates uncertainty on the part of developers regarding what
are the ground rules for an innovator in terms of permissible
designs and business models.
It is
likely that future indirect liability lawsuits will include an
intent to infringe component. Thus, it is likely that the active
inducement theory will increase the costs of litigation.21
All of a developer’s internal and external correspondences
will be analysed for signs of promoting infringement. Wu suggests
that it will be more difficult for defendants to win on summary
judgment and thus face the increased costs and uncertainty of a
trial.22
It also might encourage developers to pay off copyright owners for
peace thus making it more difficult for small innovators to survive.
Summarising,
if a developer whose behavior provides no clear evidence of intent
to induce infringement and who produces a product that is capable of
substantial non-infringing uses, then the developer should be free
of indirect liability based on a combination of the Sony and
Grokster rules. This may be the case that applies to the
pending lawsuit against the most popular peer-to-peer company
LimeWire that is widely thought to be used for direct infringement.23
It is also true that LimeWire has failed to take any steps to filter
the files shared on its network but it does employ a design that is
decentralised like Grokster, thus foreclosing actual knowledge of
infringement and lack of control over its customers. In August 2006,
copyright owners sued LimeWire for various counts of indirect
infringement.24
LimeWire has also countersued the plaintiffs for anticompetitive
activities in the online distribution of music market.25
A
remaining scenario involves a developer who provides a service with
no obvious intent to infringe, that is capable of substantial
non-infringing uses, but who may possess actual knowledge of
specific acts of direct infringement.26
The Grokster opinion left this scenario unanswered. However,
the legality of this scenario is relevant to determining the
viability of emerging competitors in the online distribution market.
QUESTION
4: What are the ground rules for indirect liability of a video
sharing web site?
Viacom’s
$1 billion lawsuit against YouTube may be decided based on
legislatively created Section 512 rules contained in the Copyright
Act and/or judicial created rules outlined above.27
YouTube utilises a centralised server client architecture
suggesting that it possesses the right and ability to control the
actions of its customers. Viacom sued YouTube for three counts of
direct infringement and three counts of indirect infringement (one
each for inducement, contributory, and vicarious liability) while
YouTube is likely to seek immunisation based on Section 512 (c),
although the law is undeveloped as to whether Section 512 can
provide a safe harbor for direct infringement liability. Viacom
seeks damages from YouTube and, more significantly, a redesign of
the technology to proactively limit or reduce the placement of
infringing content on the web site. It appears that copyright owners
view YouTube today as a piracy business similar to Napster and
Grokster, that is, an online competitor that has built its business
model on facilitating or turning a blind eye to its customers’
access to significant amounts of infringing copyrighted content.
But, with the proper redesign, YouTube could transform itself into a
legitimate distribution platform for copyrighted content as well as
user-generated content.
The
developers of the YouTube video sharing technology created a service
that its customers can use for both non-infringing and infringing
purposes. Given the relatively large volume of user-created content
uploaded to the web site, it would appear that the service is
capable of substantial non-infringing uses and thus should pass a
reasonable interpretation of the Sony test for a new dual use
technology.28
In light
of the original purpose of video sharing sites, YouTube should be
viewed as a legitimate business and not as a business intended to
facilitate predominantly copyright infringement as many viewed
Napster, Grokster, and their ilk. Thus, invoking the Grokster
rule, there appears to be little evidence of “clear expression
or other affirmative steps” taken by YouTube with the intent
to facilitate direct infringement by its customers. YouTube appears
to meet the criteria of a good-faith innovator.29
It has not promoted its service to target an audience predominantly
interested in infringing behavior and its business model is not
based on generating revenue tied to massive infringement. Although,
it does not proactively filter uploaded content (other than for
pornographic and hateful material), this, by itself, according to
the Grokster opinion, is not sufficient to suggest an act of
active inducement. Therefore, it is reasonable to conclude that
YouTube acts in good faith with no intent to induce direct
infringement. But, the interrelationship between the design of the
service and the burden of responsibility for filtering is a
continuing critical issue for innovators.30
Copyright
owners would like it to be the responsibility of video sharing sites
(and peer-to-peer networks as well) to police sites proactively for
infringing content. At first glance, this would seem to be
consistent with the approach taken in the Napster litigation. The
courts found that, given the design of its network, Napster had
control over how its customers used its service and if copyright
owners properly notified Napster of the presence of infringing files
on its index, it was deemed responsible for blocking access to those
files. It is important to recall that Napster was required to
respond to notification of infringing files and not to proactively
screen its index before such notification. In the case of YouTube,
it is unambiguous that when properly notified of the presence of
infringing content on its web site, the content is expeditiously
taken down in accordance with the procedures outlined in the Section
512 of the Copyright Act. Thus, the position that the web site must
proactively police the site for infringing content would alter the
existing burden of responsibility in the rules that technology
developers must follow.31
A complicating factor in the analysis is the assertion that YouTube
does proactively filter content if it has negotiated a licensing
agreement with a copyright owner.32
Thus, it may be possible to conduct such preemptive filtering but
it does not appear to be required by law to do so. Hopefully,
developers will continue to be free to invent new distribution
platforms without worrying about whether the innovation negatively
impacts the business models used by existing distribution channels.
Requiring developers to consider ex ante alternative designs of
their technology based on fear of litigation ex post opens up the
door to second guessing by courts and complicates the innovative
process by seemingly requiring the presence of lawyers on the
research and development teams. In turn, uncertainty about the
ground rules is likely to decrease the amount of online innovation
and solidify the position of entrenched incumbents.
The
design of new video sharing sites appears to be following the
decentralisation pattern of peer-to-peer networks. YouTube bundles
together the hosting of the uploaded video files on its own servers
and an index to locate those files. But, entrants such as
YouTVpc.com unbundled its service for it is only a video linking
site whereas content is stored at remote sites such as Dailymotion
in France and Ouou.com in China.33
Copyright owners have initiated legal action against linking sites
but have discovered that sites have re-entered the linking market
after being shutdown.
In sum,
the following rules provide a relatively certain environment for
innovators of dual use technologies that should promote competition
and the progress of science and the useful arts:
Create
a dual use technology that is capable of substantial non-infringing
uses for this technology should create significantly increased
value for society.
Create
a technology with good intent to satisfy a market associated with a
non-infringing purpose.
If
the developer of a technology is notified by copyright owners of
specific instances of infringement when and if the developer can
intervene to prevent that infringement, the developer’s
design should be adjusted to reduce or eliminate the infringing
uses without interfering with the non-infringing uses if it can be
accomplished in a cost-effective manner. This process implies that
it is the responsibility of copyright owners to monitor networks
and sites.
A
developer of a new technology is not under any obligation to design
a product/service ex ante in a particular way to accommodate the
desires of incumbent competitors.34
Rather, the design should be consistent with the existing
legislative and judicial rules as they are understood. A new
technology that merely redistributes existing value should not be
subject to liability for that reason alone.
4.2 Direct Liability Strategies
Starting
in 2003, copyright owners pursued a strategy of suing individuals
who used free peer-to-peer services alleging direct infringement of
the reproduction and presumptively distribution rights.35
Legal action against individuals is not confined to the United
States. The International Federation for the Phonographic Industry
has brought legal action against more than 3,800 individuals in
sixteen countries besides the United States.36
Copyright owners face a daunting task for it is estimated that
about 15 million United States households used unlicensed
peer-to-peer networks in 2006 to download a song.37
In any file sharing transaction, there is an uploading peer
computer (supply side) and a downloading peer computer (demand
side).38
Three economic effects could result when a consumer downloads
content from peer-to-peer networks:
(1) Downloaded content substitutes for purchases that the consumer
otherwise would have been willing to make at market prices
(substitution effect),
(2) Downloaded content represents consumption of content that the
consumer possesses a willingness to pay that is less than the market
price but above the marginal cost of distribution (consumer welfare
effect), or
(3) Downloaded content represents sampling of content that may
eventually lead to a willingness to pay for the content at market
prices as a result of experiencing the content (sampling effect).
Reason
(1) is damaging in the short run to the copyright owners’
profitability and should be the focus of a litigation strategy if
the goal of the campaign is to recapture lost sales. In addition, in
the long run, this drop in revenue may diminish the incentive to
create new content. Reason (2) represents increased consumer welfare
with no resulting loss of revenue to content owners since consumers
value the product at less than the market price but above the zero
marginal cost of distribution.39
Thus, downloading that meets this condition results in an increase
in consumer welfare for it causes a reduction in deadweight loss.
Reason (3) represents potentially increased revenue for copyright
owners and increased utility for consumers.40
A goal to
reduce infringement in general is aimed at reducing downloading for
all three reasons while a more focused goal to maximize profits
would target litigation efforts against consumers predominantly
using peer-to-peer networks for reason (1). Pursuing the latter goal
of reducing the substitution effect means that litigation efforts
would attempt to raise the expected cost of using peer-to-peer
networks or video sharing web sites for those marginal consumers
most likely to have a willingness to pay market prices. After
building a consumer choice model for file sharing and analysing the
music industry’s litigation strategy that focused its
litigation resources on high volume uploaders, Sag offers the
following demand side recommendation: “It makes much more
sense for the recording industry to target more marginal file
sharers because they are more likely to be persuaded to stop file
sharing and start buying music.” Hence, this recommendation is
reflected in the title of his article designating such presumably
marginal file sharers as twelve year-olds and grandmothers.41
Focusing
on a supply side strategy, Bhattacharjee, Gopal, Lertwachara, and
Marsden provided empirical evidence on the effect of the music
industry’s legal actions on individual file sharing behavior
in the time period 2003-04.42
The hypotheses in the study are that an event that is perceived to
increase the threat of legal action against individual file sharers
is expected to reduce the number of music files shared and to reduce
the frequency of time that an individual is online.
The
authors followed 2,056 users of the Kazaa P2P file sharing program
and examined their behavior before and after the following events:
(1) announcement of intention to pursue legal action (6/26/2003),
(2) lawsuits filed against 261 individuals (9/8/2003), (3) court
ruling against the legal process the music industry used
(12/19/2003), and (4) additional lawsuits filed using the more
complex legal process (1/21/2004). The focus of the legal action was
directed toward the supply side of the market and in particular on
those who share large amounts (referred to as substantial sharers)
of copyrighted music. The good news for the record industry is that
the empirical analysis finds general support for the two hypotheses.
Individuals reduced sharing activity and substantial sharers
decreased the number of files shared (to below the threshold/threat
level, 800 or 1,000 files shared) in response to the three negative
events for file sharers. The bad news is that there still existed a
large supply of copyrighted files. The authors conclusion is that,
“At the present time, what we can say is that the previously
substantial sharers are still tending to still actively share
(albeit fewer files), and downloading options still abound for those
seeking to download.”43
In sum, a reasonable conclusion is that the supply side strategy is
generally ineffective for it probably disproportionately directs
litigation resources on individuals unlikely to switch to paid
licensed online services and it does not significantly reduce the
ability of downloading consumers to find content, especially popular
content.
Currently,
copyright owners are focusing direct litigation attention on
university and college students who they consider as major users of
unauthorised peer-to-peer networks and who use these services as
substitution for purchasing copyrighted content. The new approach is
first to send pre-lawsuit letters to colleges and universities who
in turn are expected to notify students that copyright owners have
identified them as engaging in direct infringement. The copyright
owners offer college students discounts on settlements before
proceeding to the second step of a copyright suit. In addition, if
universities fail to employ voluntarily technology to filter their
students’ data packets, then they might lose their safe harbor
as an Internet Service Provider under Section 512 because copyright
owners threatened to go to Congress to narrow the safe harbor.44
An
advantage of direct suits compared to indirect suits is that they
attack only presumptively infringing use and do not have the
potential to reduce or eliminate non-infringing uses at the same
time. But such suits come at a considerable cost estimated to be
about $250,000 for a low-stakes case.45
Perhaps more of a long term concern to copyright owners is that the
risk of legal action against file sharers does not seem to be
changing the attitude of college students toward file sharing. It is
estimated that, although aware of the issue, 67% of college students
are not concerned with the ramifications of illegal downloading.46
5. Conclusion
Consumers
of digital content are experiencing increasing options for
accessing content online. Some of theses options are
licensed/authorised by copyright owners while others are
unlicensed/unauthorised. Unlicensed peer-to-peer networks and video
sharing sites are likely to continue to be significant challengers
to the copyright owners’ preferred/sanctioned methods of
centralised distribution for the online distribution of all forms of
digital content. The increasing decentralisation and growth of
peer-to-peer networks and video sharing web sites demonstrate that
these distribution platforms are resilient to copyright owners’
legal challenges and attempts to pollute shared file networks.
Legal
decisions centering on the rights of copyright owners influence the
path and nature of technological innovation in online distribution.
Software developers react to legal rules by attempting to redesign
online networks to insulate themselves from the costs of indirect
liability litigation. At the same time, legal decisions influence
the nature and extent of competition in the provision of online
distribution of digital content. In general, legal decisions have
resulted in technological innovation in network design becoming
increasingly decentralised. However, recent court decisions leave
unanswered key questions as to the legality and responsibilities of
certain network designs. The Viacom-YouTube litigation promises to
address some of these questions.
Copyright
owners will continue to assess the benefits and costs of alternative
legal strategies to protect their intellectual property. As the
probability of judicial wins on the indirect liability front
diminishes, it is likely that direct infringement suits targeted
against individuals that are willing to pay market prices for
digital content but instead substitute non-paying delivery services
will become more attractive.
