Content and its Discontents: What Net Neutrality Does and Doesn't Mean for Copyright

Most recent activity:

A copy of this speech is also available in
PDF format
.

By Gigi B. Sohn, President, Public Knowledge
Yale Information Society Project
Yale Law School, New Haven Connecticut

October 27, 2009

Thank you Laura and Chris for inviting me here today. It is truly an
honor to speak to you — the Information Society Project is an
important hub of scholarship and intellectual thought on some of the most
critical technology and information policy issues of the day. I am
particularly humbled to be part of the same speaker series as my friends
and colleagues Julie Cohen, Beth Noveck, Siva Vaidhyanathan and Fred Von
Lohmann.

I’m going to talk about two information policy issues today —
network neutrality and copyright enforcement — and I’ll talk
about where they do and don’t intersect. This issue really hits
Public Knowledge’s sweet spot — the organization was founded
in 2001 to ensure that citizens have access to an open communications
system at every layer. Thus, we are concerned both with copyright
holders’ control of the content layer and Internet Service
Providers’ control of the physical and applications layers of our
communications system. It has always confounded me that other media
reform groups in Washington practically run away screaming when copyright
reform is mentioned. In my opinion, we cannot have an open Internet if
large corporate copyright holders can exploit overly burdensome copyright
laws to sacrifice legitimate speech at the altar of trying to stop
piracy.

As I’m sure many of you know, net neutrality has been in the news a
lot lately. Last Thursday, the Federal Communications Commission proposed
net neutrality rules and is seeking comment on those rules for an
unusually long period of time — over four and a half months. It
will probably take several more months after that for the agency to come
up with a final order and vote on it. For those of you not familiar with
net neutrality, it is a straightforward concept. The core idea behind net
neutrality is that network providers that offer access to the Internet
should not be able to pick winners and losers among the applications,
services and content that travel over those networks. Some network
providers would like to charge application, service and content providers
for quality of service assurances; other providers want to ensure that
the applications, services and content that they own are given a
competitive advantage. But that would upset the fundamental nature of the
Internet, a principle that is built into the network’s very
architecture — that control rest at the ends of the network, and
not in the middle.

Prior to 2005, Internet access providers were prohibited from favoring or
discriminating against certain applications and content. That changed
when the Supreme Court in the Brand X case upheld an FCC
decision that found that cable-based Internet access services were
unregulated “information services.” Until that decision,
Internet access services (at least those provided by telephone companies)
were considered telecommunications services under Title II of the
Communications Act and providers were prohibited under that title from
engaging in “unjust and unreasonable” discrimination. After
the Brand X decision, the FCC quickly moved to classify
broadband wireline Internet access, broadband wireless Internet access
and broadband-over-power line Internet access as information services as
well.

In an effort to set some guidelines for network provider behavior in the
wake of the Brand X decision, the FCC in 2005 adopted four
broadband principles, “to ensure that broadband networks are widely
deployed, open, affordable and accessible to all consumers.” They
are: 1) Consumers are entitled to access the lawful Internet content of
their choice; 2) Consumers are entitled to run the applications and
access the content of their choice; 3) Consumers are entitled to connect
legal devices that do not harm the network; and 4) Consumers are entitled
to competition between networks, application, service and content
providers. But while these principles address a situation where a network
provider blocks an application, service or content, they do not address
the situation where a provider favors or degrades an application, service
or content for competitive or other reasons.

So, for the past 4+ years, Public Knowledge, along with its allies in the
public interest community and in the “edge” Internet
industry—companies like Amazon, eBay, TiVo, Sony Electronics, Skype
and Google—have been asking Congress and the FCC to reinstate the
non-discrimination rule, which was first called “open
access,” and now, thanks to Columbia Professor Tim Wu, is called
network neutrality or net neutrality.

At this point, you might find yourself asking, how does copyright fit
into this picture? Copyright issues came rather late to the net
neutrality party. At first, in the “open access” days, Disney
was an active member of the pro-open access Coalition of Broadband Users,
which also included Microsoft, Amazon, Yahoo and Apple. Disney understood
the dangers involved with being a content provider that must seek out
access to a conduit. It had for years done battle with cable operators
seeking to deny access or to extract monopoly rents.

But as movie studios increasingly lost control of their content on
digital networks, Disney and the other studios decided to focus the
majority of their policy resources on fighting infringement. Because of
the amount and value of the programming they own, the studios
didn’t fear a loss of access or discrimination much — in
fact, they are happy to pay an Internet access provider for better
quality of service or faster speeds for their content. On the other hand,
the studios began to fear that a non-discrimination rule could prohibit
network providers from enforcing their copyrights, and specifically, from
engaging in network level filtering to automatically identify and block
illegally obtained, copyrighted content. If a network provider was
prohibited from blocking or degrading applications, services and content,
how could it block unlawful transfers of copyrighted works?

Let me stop at this point to briefly explain what copyright filtering is
and how it works. At the most basic level, copyright filtering is a
method whereby network appliances use a technology known as Deep Packet
Inspection, or DPI, to inspect the data that travels over an Internet
access provider’s network, identifying content as it passes through
the filter and then either blocking or permitting that traffic based on
the instructions the filter is given. When a network provider uses DPI,
it is as if the postman grabbed your mail, opened it up, read it, and
then either sent it to its destination or stuck it in his pocket based on
the contents of your letter.

Filters can either attempt to determine the type of content that’s
in transit by looking at packet exchanges or traffic patterns —
this is what Comcast was found to have done when it illegally blocked the
Bit Torrent protocol — or by engaging in content analysis, whereby
the data in hand is compared to a registry containing the identifying
characteristics of works a copyright holder wants protected. In addition
to adding enormous costs and technological burdens to the operation of a
network, copyright filters harm end users in several ways. First, because
no filter is (or likely ever will be), smart enough to distinguish a
lawful use of copyrighted works from an unlawful one, it will necessarily
block lawful uses of content. Second, to the extent that filters use DPI,
they will have grave implications for user privacy and for lawful speech
that is completely unrelated to copyright. Recall that it was DPI that
the Iranian government used to stifle political dissent over the
country’s elections earlier this year.

If you want a more detailed explanation of copyright filters and their
policy implications, visit Public Knowledge’s homepage and download
the white paper entitled “Forcing the Net Through A Sieve: Why
Copyright Filtering is Not a Viable Solution for U.S. ISPs.”

During the last few years, the content industry has started looking for
opportunities to ensure that network neutrality does not upset their
filtering dreams. The first came during the FCC’s drafting of the
August 2008 decision declaring that Comcast’s blocking of Bit
Torrent violated the agency’s broadband principles. From their
adoption in 2005, the principles were subject to a “reasonable
network management” exception. The Comcast decision defined the
reasonable network management standard for the first time, stating that
“there must be a tight fit between [the network providers] chosen
practices and a significant goal….Its practice should further a
critically important interest and be narrowly or carefully tailored to
serve that interest.” The Commission decided that since Comcast
selectively blocked only the BitTorrent application and not other high
bandwidth applications, and because it did so at all times, and not
solely during times when its network was congested, Comcast’s
actions did not constitute reasonable network management.

But the FCC did not stop there. In dicta, it explained that it
did not wish to micromanage ISPs network management practices, and noted
that because the principles state that:

consumers are entitled to access the lawful Internet content of
their choice, network providers, consistent with federal policy, may
block transmissions of illegal content (e.g., child
pornography) or transmissions that violate copyright law.

Of course, this statement did not come out of nowhere — the studios
lobbied the FCC to include it.

So started the meme, now universally accepted, that network neutrality
only protects lawful content; and that consistent with
reasonable network management, the network provider can willfully block
or degrade both unlawful content like child pornography and unlawful
transfers
of content like the unauthorized transfer of copyrighted
works. Moreover, as the principles state — the ability of consumers
to use applications and services of their choice is subject to “the
needs of law enforcement.”

This standard gives copyright holders a wide berth with which to enforce
their rights online. Nothing in the principles interferes with the notice
and takedown requirements of the Digital Millennium Copyright Act (DMCA).
Nothing prohibits filtering at the edge, like YouTube’s automated
content identification system. Nothing prohibits a network provider from
sending warning notices to subscribers engaging in unlawful conduct,
which, according to a Disney lobbyist, effectively stops illegal behavior
80% of the time. Nothing prohibits the government or copyright holders
from going to court and bringing criminal and/or civil actions against
illicit websites and applications or, as they often do, from stopping
illegal behavior using the mere threat of excessive statutory damages.

But how does network level copyright filtering, which necessarily blocks
both lawful and unlawful content, fit into the reasonable network
management standard? While Hollywood and the recording industry claim
that such filters work 98% of the time, they (as usual) have no empirical
evidence to back up that claim, and even if they did, that 2% of errors
could contain the next great “Macaca” moment that changes the
course of an election. Of course, I am referring to the 2006 video that
showed then Senator George Allen making a racially offensive remark to a
videographer from his opponent’s campaign. And what should be made
of the fact that copyright filtering does nothing to ensure the smooth
operation of the network (in fact, it accomplishes the opposite) and
therefore is not network management but content
management?

So let’s look again at the Comcast reasonable network management
standard:

there must be a tight fit between [the network provider’s] chosen
practices and a significant goal….Its practice should further a
critically important interest and be narrowly or carefully tailored to
serve that interest.

It seems to me that even if we assume that copyright filtering is network
management and not content management, there is little doubt that
copyright filtering cannot meet that standard. There is no “tight
fit” between the practices and the goal of stopping infringement.
And even if stopping or slowing infringing behavior can be considered a
“critically important interest,” the FCC would be hard
pressed to find that filtering, which is a blunt instrument, is
“carefully tailored to serve that interest.”

I don’t doubt that Hollywood agrees with me (for once), because
ever since the Comcast decision, it has used every opportunity to have
Congress and the FCC to declare flat out that copyright filtering is
“reasonable network management,” thereby removing any
uncertainty as to whether such filtering would meet the Comcast standard.
The first effort came during Congress’ consideration of the
American Recovery and Reinvestment Act, known more commonly as the
economic stimulus bill. Senator Dianne Feinstein proposed an amendment to
the portions of the bill that empowered the National Telecommunications
and Information Administration of the Department of Commerce to give
grants totaling $4.6 billion dollars for projects that would promote
broadband deployment and adoption. The bill required that grantees
operate their networks in a non-discriminatory manner. Seizing on that
openness requirement, the Feinstein amendment required the NTIA to:

…allow for reasonable network management practices such as
deterring unlawful activity, including child pornography and copyright
infringement.

Copyright filtering certainly “deters” copyright
infringement, and as such would almost certainly be considered reasonable
network management under the Feinstein definition.

Public Knowledge spearheaded the opposition to this language, and with
the aid of our public interest and industry allies, the language was
dropped in exchange for a pointed question to then FCC Chairman-designate
Julius Genachowski about net neutrality and copyright infringement. While
the NTIA was developing grant guidelines pursuant to the stimulus bill,
Hollywood again tried and failed to condition the non-discrimination
requirement with a reasonable network management exception encompassing
efforts to “deter” copyright infringement.

The economic stimulus bill also required the FCC to develop a National
Broadband Plan — that is, a roadmap for how this country can first
achieve a universally accessible and affordable broadband system that
revolutionizes not only communications, but also education, health care,
energy independence, etc. and second, how it can encourage those who
haven’t yet adopted broadband to do so. Interestingly, FCC
officials have made clear that net neutrality will not be discussed in
the context of the broadband plan, leaving it to the rulemaking I
discussed earlier. But that has not deterred Hollywood from advancing its
agenda one iota. The studios argue that protection of copyright must be
part of the National Broadband Plan because without such protection, they
will not make valuable content available online, and without their
valuable content online, people who do not already subscribe to broadband
will not do so. While popular Hollywood content is undoubtedly one driver
of broadband adoption, it is hardly the only, or even the most important
driver of adoption. Indeed, social networks like Facebook are arguably
more important drivers of broadband adoption, and user generated content
appears to be more popular than Hollywood fare on websites like YouTube.
While I have received assurances from FCC staff that copyright issues
will not be addressed in the broadband plan, Hollywood’s relentless
lobbying of the FCC could certainly change that calculus.

This brings us back to the net neutrality rulemaking, which may be
Hollywood’s best opportunity to get explicit permission to pressure
network providers to filter. Remember that I said earlier that the
Comcast reasonable network management standard is too stringent to permit
filtering. So if you can’t get an explicit law or rule that
clarifies that filtering is indeed reasonable network management,
what’s the next best thing? Trying to lower the reasonable network
management standard.

In a move that must have had corks popping in Los Angeles, the FCC is
seeking comment on that very issue. In paragraph 137 of its notice of
proposed rulemaking, the Commission tentatively concludes that the
Comcast reasonable network management standard is “unnecessarily
restrictive in the context of a rule that generally prohibits
discrimination subject to a flexible category of reasonable network
management.” In other words, the Commission is arguing that a
strict reasonable network management standard may be unnecessary if the
Commission adopts non-discrimination requirement, and it seeks comment on
that proposal.

The FCC proposes the following definition of reasonable network
management to replace the Comcast standard:

Reasonable network management consists of (a) reasonable practices
employed by a provider of broadband Internet access service to (i)
reduce or mitigate the effects of congestion on its network or to
address quality-of-service concerns; (ii) address traffic that is
unwanted by users or harmful; (iii) prevent the transfer of unlawful
content; or (iv) prevent the unlawful transfer of content and (b) other
reasonable network management practices.

By eliminating the requirement that the network provider demonstrate that
its management practices bear even the slightest relation to an important
network management goal, the FCC has greatly weakened the standard.
Moreover, the proposed definition is circular: reasonable network
management consists of “reasonable network management
practices?” Maybe I’m missing something, but huh?

I must say also that I don’t quite understand the FCC’s
rationale for weakening the standard. The fact that they are reinstating
a non-discrimination requirement means very little if the exception to
the requirement is so broad as to undermine it.

In any event, this definition appears to open the door wide for
network-level copyright filtering. The FCC need only determine that the
technique was a “reasonable practice” to “prevent the
unlawful transfer of content.” A network provider needn’t
show that its network management practices are a “tight fit”
or “narrowly tailored” to prevent the unlawful transfer of
content. In other words, under this standard, a 98% accuracy rate, or
maybe even 51%, would likely be considered reasonable.

Now before Hollywood lobbyists start drinking that champagne, there is
other language in the notice that could be interpreted to cabin the
reasonable network management standard as it applies to copyright
filtering. In paragraph 139 of the order, the FCC again emphasizes that:

[o]pen Internet principles apply only to lawful transfers of content.
They do not, for example, apply to activities such as unlawful
distribution of copyrighted works, which has adverse consequences on the
economy and the overall broadband ecosystem. In order for network
openness obligations and appropriate enforcement of copyright laws to
co-exist, it appears reasonable for a broadband Internet access service
provider to refuse to transmit copyrighted material if the transfer of
that material would violate applicable laws. Such a rule would be
consistent with the Comcast Network Management Practices Order,
in which the Commission stated that “providers, consistent with
federal policy, may block…transmissions that violate copyright
law.”

In this language, there is a lot of emphasis on unlawful transfers and
the ability to block unlawful transfers of content if the transfer of the
material would violate copyright laws. But those qualifiers are not in
the proposed rules themselves — the rules are purposefully
barebones, and include only the proposed new definition of reasonable
network management.

I’m hoping that my Public Knowledge colleague Art Brodsky is
correct in saying that the notice is Talmudic in nature — the
proposed regulations are spare, but there is a lot of language in the
notice that will help to inform this FCC and future FCCs when the time
for interpretation arrives. But it is not due to concerns about copyright
that Public Knowledge will seek to tighten the FCC’s proposed
standard for reasonable network management. The validity and
enforceability of the rules rest on that definition, and while we
certainly agree that network providers should be able to ensure the
proper and secure operation of their networks, and even that wireless
providers should have a bit more leeway (at least at first) because of
bandwidth constraints and technological differences from wireline
services, we can’t support an exception that swallows the rule.

So the next 6 months or so are sure to be fascinating on many different
levels. Given its twin goals of open networks and balanced copyright,
Public Knowledge in particular is in an interesting and somewhat
difficult position — wanting very much to support the FCC in its
adoption of enforceable net neutrality rules, but very much concerned
about what the proceeding may mean for its copyright policy goals. It is
a much easier calculus for a company like Verizon, which opposes
copyright filtering and net neutrality regulations. At Public
Knowledge’s bi-weekly copyright reform meeting, I asked a Verizon
lawyer how the company was going to address the issue. She said, and I
paraphrase — “it’s really hard, but this proceeding may
drive Internet service providers and Hollywood into each other’s
arms,” because they both want a broad network management exception.

Ultimately, I agree with Chairman Genachowski, who said in his September
21 speech announcing the commencement of the rulemaking that “the
enforcement of copyright and other laws and the obligations of network
openness can and must co-exist.” But the question as to what
constitutes “appropriate” or “reasonable”
copyright enforcement by network providers will inevitably challenge this
goal. I welcome all of you to help inform the debate — the
Commission really wants to hear from people outside the Beltway and I
know that they will make good on their promise to make this process open,
transparent and welcoming to your input. Public Knowledge and its allies
need your help to ensure that the FCC’s open Internet rules promote
their intended purpose — to promote creativity, innovation, civic
discourse and economic growth. Thank you.