Network neutrality article mischaracterizes Canadian incidentBarbara's response --

[David Farber <dave () farber net>]

Begin forwarded message:

From: Barbara van Schewick <schewick () stanford edu>
Date: February 22, 2009 2:05:01 PM EST
To: David Farber <dave () farber net>
Subject: Re: Network neutrality article mischaracterizes Canadian  
incident

In his e-mail, Brett Glass critiques a portion of my article on  
network neutrality that was recently published in the Communications  
of the ACM as part of a point/counterpoint exchange with David Farber (http://doi.acm.org/10.1145/1461928.1461942 
).

In the paragraph that Mr. Glass focuses on, I argue that “network  
providers may have an incentive to block unwanted content that  
threatens the company’s interests or does not comply with the network  
provider’s chosen content policy.” (CACM article, p. 32) I use three  
recent incidents of content-based discrimination to illustrate this  
claim.

Mr. Glass claims that my necessarily brief descriptions of the  
incidents mischaracterize what actually happened, and that the  
incidents, once characterized correctly, do not support an argument in  
favor of network neutrality regulation. I don’t agree.

According to Mr. Glass, my article claims “that Telus, a Canadian  
telephone company and ISP, blocked a union Web site during a labor  
dispute.” Mr. Glass then reveals that, when he researched this  
incident, he “learned […] that the site that was blocked by Telus was  
not that of the union. It was called "Voices for Change," and was run  
by a union member named David DiMaria.”

I don’t disagree with the facts Mr. Glass discovered, but he does not  
accurately describe my argument. My article does not claim that the  
website was the union’s website, but clearly states that the web site  
was run by a member of the Telecommunications Workers Union: “In 2005,  
Telus, Canada’s second largest ISP, blocked access to a Web site that  
was run by a member of the Telecommunications Workers Union. At the  
time, Telus and the union were engaged in a contentious labor dispute,  
and the Web site allowed union members to discuss strategies during  
the strike.” (CACM article, p. 32)

According to press reports at the time, Telus justified the blocking  
as follows: “[T]he company said the site suggested striking workers  
jam Telus phone lines, and posted pictures of employees crossing the  
union picket lines. Telus spokesman Drew Mcarthur said advocating  
jamming lines hurt the company, and access to the pictures threatened  
the privacy and safety of employees.” (CBC News, http://www.cbc.ca/canada/story/2005/07/24/telus-sites050724.html) 
 Mr. Glass cites from a letter that DiMaria’s lawyer wrote to DiMaria  
in connection with a settlement between Telus and DiMaria, in which  
the lawyer states that some of the content on the website would indeed  
have justified an injunction; the lawyer explicitly describes two  
posts that identify and threaten employees who continue to work during  
the strike. As part of the settlement, DiMaria agreed to remove  
content (including photographs), posted after a specific date “with  
the intent of intimidating or threatening Telus’ employees.”

According to Mr. Glass, these facts indicate that the blocking of the  
website was justified as an attempt by Telus “to prevent its employees  
from suffering physical harm.”

I don’t agree. While the facts indicate that the site included  
threatening content that needed to be removed, there are very clear  
legal procedures (such as getting a court order to get the content  
removed) for realizing this goal. Every person or entity which is  
threatened by online information needs to follow these procedures to  
protect itself. There is no reason why a network provider should have  
different rights, just because it is technically able to engage in  
some sort of self-help.

In fact, the Telus case highlights some of the reasons why network  
providers should not be allowed to engage in self-help through blocking.

First, compared to an injunction, blocking is both over- and  
underinclusive. According to investigations by the OpenNet Initiative  
at the time of the blocking, by banning access to the IP address of  
the server on which the website was hosted, Telus not only blocked  
access to “Voices for Change”, but also to 766 other websites hosted  
on the same server, among them an engineering company, a breast cancer  
fundraising site, or a Colorado based electronic recycling company. (http://opennet.net/bulletins/010 
)

With respect to the website itself, the blocking did not only affect  
the illegal content, but also the many legitimate discussions that  
were taking place on the site (according to the lawyer’s letter to  
DiMaria, “the site was hugely popular; 39,000 hits in a month. We  
assumed that it was very effective as a forum for strike  
communication; and comments critical of company during strike.” (http://flickr.com/photos/penmachine/29590389/ 
, p. 2 of the letter)

At the same time, the blocking wasn’t an effective way of preventing  
access to the illegal information: it only prevented customers of  
Telus’ Internet service (and customers of ISPs’ who were downstream  
from Telus) from accessing the site directly. Even Telus’ customers  
were still able to get access through a proxy site or one of the  
mirror sites that had been set up.

Second, the relevant legal rules assign the right to decide which  
content is illegal to a court which has no interest in the matter. As  
the Telus’ case shows, an entity whose own interests are at stake may  
have an incentive to interpret more liberally what needs to be  
removed. As indicated above, Telus consistently justified the blocking  
not just with the threat to employee’s safety, but also with the  
online discussions about jamming Telus’ customer service lines that  
were taking place on the site. According to Telus spokesman Drew  
Mcarthur, “advocating jamming lines hurt the company.” By contrast,  
the union argued that “union members who post online exchanges about  
jamming Telus service lines are acting within their rights.”
 (CBC News, http://www.cbc.ca/canada/story/2005/07/24/telus-sites050724.html) 
. These additional concerns related to Telus’ business interests  
featured prominently in other statements by the company: for example,  
in a standardized letter to customers who complained about the  
blocking, Telus wrote that the website was blocked, “because it  
publishes confidential TELUS documents, photos of TELUS team members  
who have chosen to continue to work, and instructions on how to carry  
out harmful actions that impede TELUS’ ability to serve our  
customers.”, and that Telus “blocked access to the site to protect our  
employees, our assets, and reduce activities that are clearly designed  
to limit our ability to provide the highest level of customer service  
possible.” http://www.penmachine.com/2005/07/continuing-telus-website-blocking-saga.html) 
. However, neither the settlement with DiMaria, nor in the injunction  
against the Telecommunications Workers Union that Telus obtained in  
court address these business-related concerns. While there may be  
other reasons, this suggests that these concerns were unable to  
withstand judicial scrutiny.

All this suggests that while there clearly was some illegal content on  
the web site, blocking the site was not the right method to get it  
removed. In addition, Telus’ public statements regarding their other  
motivations that go beyond the safety of their employees suggest that  
using the Telus story as an example of a company that “blocks unwanted  
content that threatens the company’s interest” is justified.

According to Mr. Glass, my “article similarly mischaracterizes two  
other incidents. In one, a Verizon Wireless employee -- in violation  
of company policy -- delayed issuing an SMS "short code" to Naral.  
Contrary to the assertions in the article, this incident had nothing  
to do with the Internet and was the result of an employee's misconduct  
rather than a problematic company policy.”

I don’t agree with this characterization of what actually happened.  
Here is how the New York Times described the incident in a front page  
article on September 27, 2007 (http://www.nytimes.com/2007/09/27/us/27verizon.html 
), drawing on written copies of Naral’s communication with Verizon and  
an interview with a Verizon spokesperson:

“Saying it had the right to block ‘controversial or unsavory’ text  
messages, Verizon Wireless has rejected a request from Naral Pro- 
Choice America, the abortion rights group, to make Verizon’s mobile  
network available for a text-message program. […]
In turning down the program, Verizon, one of the nation’s two largest  
wireless carriers, told Naral that it does not accept programs from  
any group ‘that seeks to promote an agenda or distribute content that,  
in its discretion, may be seen as controversial or unsavory to any of  
our users.’ Naral provided copies of its communications with Verizon  
to The New York Times. […] A spokesman for Verizon said the decision  
turned on the subject matter of the messages and not on Naral’s  
position on abortion. ‘Our internal policy is in fact neutral on the  
position,’ said the spokesman, Jeffrey Nelson. ‘It is the topic  
itself’ — abortion — ‘that has been on our list.’” (http://www.nytimes.com/2007/09/27/us/27verizon.html 
)

In an article the next day, the New York Times added that “In turning  
down the request last week, Verizon told Naral that it ‘does not  
accept issue-oriented (abortion, war, etc.) programs — only basic,  
general politician-related programs (Mitt Romney, Hillary Clinton,  
etc.).’” (http://www.nytimes.com/2007/09/28/business/28verizon.html)

Thus, contrary to Mr. Glass’ claims, Verizon rejected (and not only  
“delayed”) issuing an SMS short code to Naral based on an existing  
content policy. This decision was not the “result of an employee’s  
misconduct” “in violation of company policy”. When the New York Times  
contacted Verizon to investigate the story, the company officially  
confirmed the decision through a spokesperson, who justified the  
decision by reference to the content policy. Given these facts, I  
think it is legitimate to use this incident as an example of a network  
provider who blocks content that does not comply with the network  
provider’s chosen content policy.

After the incident had made the front page of the New York Times and  
caused a public outcry, Verizon changed its mind and reversed its  
original decision (a development my article mentions in footnote d) on  
p. 32). In a statement following the reversal, Jeffrey Nelson, the  
same spokesperson who, a few days ago, had defended Verizon’s  
rejection of Naral’s request based on Verizon’s existing content  
policy, now framed the incident as “an incorrect interpretation of a  
dusty internal policy” that “was designed to ward against  
communications such as anonymous hate messaging and adult materials  
sent to children” and that had been developed “before text messaging  
protections such as spam filters adequately protected customers from  
unwanted messages.” (Statement issued by Jeffrey Nelson, as cited by  
the New York Times, http://www.nytimes.com/2007/09/28/business/28verizon.html) 
 Given Verizon’s public statements immediately after the rejection,  
these subsequent attempts to control the damage by re-interpreting  
what happened do not change the interpretation of the original  
incident that my article describes.

In addition to the Verizon/Naral story, my article points to a second  
case in which a company deleted content based on the company’s content  
policy. In this incident,
“AT&T deleted words from a Webcast of a Pearl Jam concert in which the  
singer criticized George W. Bush” (CACM article, p. 32). The deleted  
words were “George Bush, leave the World alone” and “George Bush, find  
yourself another home.” (InformationWeek, http://www.informationweek.com/story/showArticle.jhtml?articleID=201310731 
 ).

According to Mr. Glass, my article mischaracterizes this incident,  
because “AT&T received the broadcast in this form from a third party  
and was not responsible for the editing.”

This does not accurately describe what happened. AT&T streams the web  
casts through a streaming portal called Blue Room. Since the website  
is open to all age groups, AT&T tries to remove “excessive profanity”  
and “nudity” from the web casts. In an interview with InformationWeek  
following the blocking, AT&T spokesman Michael Coe described the  
practice as follows: "We work with an agency to work with festival to  
secure rights for broadcast," he said. "This is obviously not a core  
competency for A&T. They hire Web monitors. The live concerts, when  
streamed into the Blue Room, are not age-restricted, meaning anyone of  
any age can enter the Blue Room. They are told only to edit for  
excessive profanity, not really for the songs, but the banter going on  
between band members or band members and the audience, as well as any  
nudity that can arise."
http://www.informationweek.com/story/showArticle.jhtml?articleID=201310731

Thus, there was a content policy (i.e. “edit for excessive profanity  
and nudity”) established by AT&T, and this content policy was enforced  
by “Web monitors.” The fact that these “Web monitors” were hired by an  
agency acting on behalf of AT&T, not by AT&T itself, does not really  
make a difference. They were still acting on behalf of (and following  
instructions by) AT&T.

After the blocking had happened, AT&T immediately characterized the  
incident as a “mistake.” “We regret the error”, a spokesperson said in  
an interview with the LA Times technology blog. ”There was no  
profanity. It was a mistake.” (http://opinion.latimes.com/bitplayer/2007/08/att-drops-pearl.html 
) Thus, AT&T did not deny the fact that there was a content policy,  
but claimed that the content policy had been applied in the wrong way  
(because it was a political statement, not profanity).

Does this affect the relevance of the incident for the network  
neutrality debate? I don’t think so. In a world without network  
neutrality rules, network providers are free to adopt content policies  
such as the ones described above. This may create two types of  
problems: First, the content policies themselves may prevent end users  
from accessing certain content and affected content producers from  
reaching their audience, as in the Verizon example. In addition,  
enforcement of the content policy may result in (deliberate or  
inadvertent) misclassifications, resulting in the exclusion of  
additional content. In other words, mistakes in enforcing the content  
policy are a foreseeable consequence of a regime that allows network  
providers to pursue content policies in the first place.

At the same time, the impact of content blocking on Internet service  
customers’ ability to listen or speak is the same, regardless of  
whether the blocking is based on a correct or incorrect application of  
the content policy: “If I am interested in content that my network  
provider has chosen to restrict, my ability to educate myself,  
contribute to a discussion on this subject, and make informed  
decisions is impeded. Instead, ISPs gain the power to shape public  
discourse based on their own interests and idiosyncratic content  
policies.” (CACM article, p. 32) Thus, by showing that  
misclassifications do indeed happen, the AT&T case illustrates one of  
the potential harmful consequences of content policies.

Finally, Mr. Glass argues that the last two examples (Verizon/AT&T)  
are irrelevant for the network neutrality debate, because neither  
Verizon nor AT&T were acting as ISPs. This argument is explicitly  
addressed in my article: “While the latter two examples are not direct  
examples of ISPs restricting content on their networks (Verizon  
Wireless restricted a service on its wireless mobile network, not the  
wireless Internet, while AT&T acted in its role as a content provider,  
not as ISP), it is easy to imagine virtually identical incidents in  
which an ISP enacts a content policy and restricts content on its  
network accordingly.” (CACM article, p. 32) Thus, while the actual  
incidents happened in a slightly different context, they illustrate  
the type of blocking that network neutrality proponents are concerned  
about.

Barbara van Schewick
--
Barbara van Schewick
Assistant Professor of Law and (by Courtesy) Electrical Engineering
Co-Director, Center for Internet and Society
Stanford Law School

Crown Quadrangle
559 Nathan Abbott Way
Stanford, CA 94305-8610

Phone:  650-723 8340
E-Mail: schewick () stanford edu

----------------------


Begin forwarded message:

From: Brett Glass <brett () lariat net>
Date: February 18, 2009 10:49:34 PM EST
To: "Dave Farber" <dave () farber net>, "Ip ip" <ip () v2 listbox com>
Subject: Network neutrality article mischaracterizes Canadian incident

Dave, and everyone:

As you may recall from an article referenced on this list not long
ago, Barbara van Schewick claimed, in a recent CACM article in which
she claimed that there was an urgent need for "network neutrality"
regulation, that Telus, a Canadian telephone company and ISP, blocked
a union Web site during a labor dispute. I've researched this claim,
and have discovered that Ms. van Schewick was not telling the whole
story.

The first thing I learned was that the site that was blocked by Telus
was not that of the union. It was called "Voices for Change," and was
run by a union member named David DiMaria. Among other things, the
site contained pictures of non-union employees who were crossing the
picket lines, together with their home addresses, their phone numbers,
and threatening language advocating that union members physically harm
these "scabs." The site also instructed visitors to jam Telus'
customer support lines, and was thus inciting another illegal activity.

What did Telus do? Seeing that its employees could well be harmed, it
did what it could to protect their physical safety. It blocked the Web
site immediately. And since it couldn't block the site completely (it
could only keep its own customers from going to it), Telus then went
to court asking for an injunction requiring the threats to be taken
down.

Ultimately, the case was settled out of court. A copy of the
settlement between Telus and the owner of the Web site is available on
Flikr at

http://flickr.com/photos/penmachine/29590389/

(This is a very large GIF file -- not the ideal format for such a
document. But if you zoom in, it is readable and tells the story.)

A letter from his lawyer (the first two pages) advises Mr. DiMaria
that there was indeed threatening material on the site that would have
justified an injunction. (The lawyer describes it as "thoroughly
reprehensible.") It doesn't enumerate everything that was posted, but
mentions two examples. One of the employees who was threatened -- a
dark skinned East Indian -- was described as a "terrorist" and his
name, address, and phone number were published. Due to the threats, he
was compelled to pack up his family and leave town.

And this is, as Paul Harvey would say, "the rest of the story." Was
Telus justified in attempting to prevent its employees from suffering
physical harm? I believe so, and would have done precisely what Telus
opted to do were I in the same situation.

Ms. van Schewick's article similarly mischaracterizes two other
incidents. In one, a Verizon Wireless employee -- in violation of
company policy -- delayed issuing an SMS "short code" to NARAL.
Contrary to the assertions in the article, this incident had nothing
to do with the Internet and was the result of an employee's misconduct
rather than a problematic company policy. In the other, a Webcast
carried by AT&T had certain political remarks deleted. However, AT&T
received the broadcast in this form from a third party and was not
responsible for the editing.

In short, in none of these cases does it appear that an ISP was
involved in unwarranted censorship of political content.

In the course of heated debates -- such as the ongoing one regarding
"network neutrality" (AKA regulation of the Internet) -- we must
always be skeptical of, and be careful to verify, the assertions made
by both sides rather than taking them at face value. And in this case,
it's especially important, because the dangers of getting it wrong --
in this case, potentially regulating the Internet in a way that will
harm innovation, competition, and availability -- are so great.

--Brett Glass








-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com