Private Regulation and Public Policy:
Toward Effective Restriction
of Internet Hate Propaganda
Jane Bailey*
freedom of expression
l’approche
La propagande haineuse sur Internet ravive le ddbat
entre des conceptions divergentes de la ddnocratie et de la
libert6 d’expression. A l’idre d’un mareh d’ides non
r~glementd, soutenue aux ttats-Unis par la jurisprudence
sur le premier amendement de la constitution de ce pays,
s’oppose
canadienne. Au Canada,
la
constitution confere aux autorit~s gouvernementales
le
pouvoir de limiter l’exercice de la libert6 d’expression IA oil
le march6 ne le fait pas. En d-pit de la relative facilit6 avec
laquelle la propagande haineuse peut dlire domicile sur des
serveurs amriecains protdges par le premier amendement, la
r~glementation publique nationale et intemationale n’est
toujours pas inutile. Les restrictions de nature priv~e
imposes A la propagande haineuse peuvent contribuer A
faire obstacle aux lacunes resultant de l’hegdimonie du
premier amendement, mais ne sauraient se substituer A la
protection d’engagements nationaux et internationaux en
faveur de l’6galitd et de la diversit6. L’auteur sugg&e qu’A
ces inconvrnients pratiques s’ajoute l’incapacitd damontrde
du marchd A amrliorer
la condition de groupes
historiquement ddsavantagds. II serait possible de rem&lier
A certains de ces inconvdnients, dans la mesure oft
la
rdglementation publique puisse guider et surveiller les
mrcanismes privrs de contr6le, cr6ant une certaine
transparence et un certain degr6 de responsabilit6. L’auteur
conclut qu’A long terme, la creation et la surveillance d’un
code de conduite pour les foumisseurs de services Internet
(ISP) par une organisation intemationale respectre telle que
les Nations Unies pourrait permettre de guider l’activit6
privee de mani-e A cc qu’elle puisse servir de manirre plus
systematique des objectifs publics liss aux droits de la
personne.
Internet hate propaganda revives debate regarding
competing visions of
and
democracy-pitting the unregulated marketplace of ideas
vaunted in US First Amendment jurisprudence on racist
speech against approaches such as Canada’s, which
envision a role for the state
in limiting de-liberating
exercises of private power unchecked by the marketplace
itself. Despite the relative ease with which Internet hate
propagandists may shift the “location” of their message to
First Amendment-protected servers in the US, national and
international public regulation of hate propaganda have not
outlived their usefulness. Existing private restriction of hate
propaganda may assist
in resisting First Amendment
hegemony, but is not an adequate substitute for protecting
fundamental national and international commitments
to
equality and diversity. The practical drawbacks of many
private enforcement mechanisms are compounded by
policy concerns underlain by a record of the private
market’s
the conditions of
historically disadvantaged groups. Certain of
these
drawbacks might be alleviated
through continuing
expressions of policy pursuant to public regulation aimed at
guiding private regulation and imbuing it with a degree of
transparency and accountability. In the long term, creation
and monitoring of an Internet Service Provider (“ISP”)
model code of conduct by a respected international body,
such as
in more
systematically harnessing private action in service of public
human rights objectives.
the United Nations, might assist
inadequacy in ameliorating
*Assistant Professor, University of Ottawa Faculty of Law, Common Law Section,
jbailey@uottowa.ca. Thanks to The Centre for Innovation Law and Policy and the Ontario Graduate
Scholarship program for funding support, Shawn Pudsey for his interest and research assistance,
Professor Trevor Farrow of the University of Alberta and two external reviewers for their thoughtful
comments on earlier drafts. All shortcomings, however, remain those of the author. For purposes of
full disclosure, readers should be aware that the author assisted as co-counsel for the complainant
Sabina Citron in Citron, infra note 10.
McGill Law Journal 2003
Revue de droit de McGill 2003
To be cited as: (2003) 49 McGill L.J. 59
Mode de r6fdrence : (2003) 49 R.D. McGill 59
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
Introduction
1.
Internet Hate Propaganda: The Scope of the Problem
II. Canadian Approach to Hate Propaganda
A. Amendment of the CHRA and the Code
1. CHRA Amendment
2. Code Provisions and Amendment
B. Canadian Constitutional Review of Hate Propaganda
Restrictions
III. First Amendment Protection of Hate Propaganda
IV. Conflicting Democratic Visions
V. First Amendment Limitations on State-Based International
Regulation of Internet Hate Propaganda
A. Procedural Agreements Calling for State Action: Draft
Convention on Jurisdiction and Foreign Judgments in Civil
and Commercial Matters
International Agreements Harmonizing Substantive Law:
The Additional Protocol to the Cybercrime Convention
B.
VI. Emerging Private Regulation
A. Privately Implemented Technological Solutions
1. Filtering
2. Zoning
B. Acceptable Use Policies
1. The America Online Model
2. Limitations on Effective Restriction Through AUPs
C. US-Based Self-Regulatory Organizations
1. ISP Organizations in the United States
2. Limitations on Effective Enforcement Through United
States ISPAs
D. Ad Hoc US ISP Responses to Extra-territorial Public Policy
1. “Pledging” Proactive Observance of “Local” Laws
2. Private Enforcement in Response to Public Decisions
E. Private Regulation Is Not a Substitute for Public Regulation
1. Practical Limits
2. Policy Issues
62
63
66
66
67
68
69
72
74
76
77
78
80
80
81
83
84
85
87
88
88
90
90
91
91
93
93
94
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
VII. Organizing Private Action to Work Toward Public Goals
A. Goals for “Publicizing” Private Action
B. Advancing the Project
1. Who Might Be of Assistance?
2. What Steps Should Be Taken?
C. Shortcomings
Conclusion
61
97
97
97
97
101
101
102
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
We have exported to the world, through the architecture of
the Internet, a First Amendment in code more powerful than
our own First Amendment in law.
Lawrence Lessig’
the protection of
Americans take free speech a bit more seriously than the
Brits, the French, the Germans and the rest of the world.
And, yes, America could become the guardian of free speech
worldwide by offering
the First
Amendment over the Net to millions of people who have
been denied the right to speak freely in their own countries.
Adam D. Thierer, CATO Institute2
Under the First Amendment the government must leave to
the people
the evaluation of
Totalitarian
governments
today rule much of the planet, practicing
suppression of billions and spreading dogma that may
enslave others. One of the things that separates our society
from theirs is our absolute right to propagate opinions that
the government finds wrong or even hateful.
ideas …
The ideas of the Klan may be propagated … The Nazi
Party may march
through a city with a large Jewish
population…. People may seek to repeal laws guaranteeing
equal opportunity
the
constitutional amendments granting the vote to blacks and
women. They may do this because “above all else, the First
Amendment means that government has no power to restrict
expression because of its message [or] its ideas…”
in employment or
to revoke
American Booksellers Association v. Hudnut3
Introduction
The relatively low cost and the dissemination power of Internet communication
make it an increasingly preferred medium for hate propagandists. Canada, other
nations, and international collectives have developed or modified existing public
this growing problem. Unfortunately, First Amendment4
regulation to address
protection of the growing tide of hate propaganda emanating from servers located in
1 Lawrence Lessig, Code and Other Laws of Cyberspace (New York: Basic Books, 1999) at 167
[emphasis in original].
2 Adam D. Thierer, “Web Resrictions Unlikely to Muzzle Neo-Nazi Speech” (15 January 2001),
online: Cato Institute
‘ 771 F.2d 323 (7th Cir. 1985) at 328, 106 S. Ct. 1172 [Hudnut cited to F.2d]. While this quote
undoubtedly overstates the degree to which US First Amendment jurisprudence forbids government
restriction of expression generally, this kind of language often surfaces in the context of restrictions on
hate propaganda and pornography.
4 U.S. Const. amend. I.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
63
the United States challenges the ability of other nations to pursue alternative public
policies, such as the one underlying recent Canadian legislative amendments enacted
to clarify the application of restrictions on Internet hate propaganda. The export of the
First Amendment approach to Canada is not properly characterized as an export of
free expression. The Canadian Charter of Rights and Freedoms protects free
expression and does so on the basis of a comprehensive conception of both public and
private forces affecting individual liberty, in accordance with international human
rights commitments.
The First Amendment challenge to restricting hate propaganda emanating from
within the US suggests that technology has overtaken public regulation at both
national and international levels. Given that private action is generally not subject to
US constitutional scrutiny, emerging regulation by private actors in the US presents
an opportunity to restrict the cross-border flow of hate propaganda. For both practical
and policy reasons, private regulation
inadequate substitute for public
regulation, particularly in the area of human rights. Nevertheless, ongoing ad hoc
private efforts to restrict hate propaganda and other illegal and offensive content invite
consideration of the ways in which public initiatives may be used to encourage more
systematic private efforts while at the same time imbuing them with a degree of public
accountability.
is an
This paper explores these issues in seven parts. Part I describes the scope of the
Internet hate propaganda problem. Part II examines the Canadian approach to hate
propaganda, focusing on legislative amendments designed to specifically address the
Internet. Part HI contrasts Canadian and US approaches, demonstrating the First
Amendment challenge to restricting the cross border flow of hate propaganda
emanating from the US. Part IV briefly discusses the different underlying democratic
visions animating the two approaches, arguing that the US approach has no prior
claim in furthering the interests of democracy. Part V explores the degree to which the
First Amendment undercuts the efficacy of public international responses based on
state action, suggesting an
is
necessary. Part VI identifies and discusses practical and policy limitations to some
current private on-line content regulation. Part VII comments on the prospect for an
internationally facilitated strategy that seeks to integrate more publicly accountable
and systematic private regulation with vigilant enforcement of public regulation.
investigation of private regulatory opportunities
I.
Internet Hate Propaganda: The Scope of the Problem
Hate propaganda6 is disseminated on the Internet using a variety of applications,
from Web sites to newsgroups to e-mail to on-line games. The exact scope of the
5 Part I of the Constitution Act, 1982 being Schedule B to the Canada Act 1982 (U.K.) 1982, c. 11
[Charter].
6 Unless otherwise defined (e.g., as defined in legislation), the term “hate propaganda” describes
messages aimed at; or with the effect of, inciting hatred or contempt for individuals or groups of
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
problem remains unclear. While there has been some effort to collect and analyze
Internet hate propaganda based on race, ethnicity, and national identity (“racist
propaganda”), until recently there has been less focus on hate propaganda based on
other grounds, such as gender and sexual identity. Nevertheless, the available reports
suggest widespread and growing use of the Internet to disseminate messages of hate
against groups identifiable on all of these grounds, as well as against the individual
members of those groups.
Reports suggest that racist Web sites alone have grown steadily from a single
white supremacist Web site in 1995′ to a reported 4,000 racist Web sites in 2001, an
estimated 2,500 of which emanate from US servers.8 Groups such as the Ku Klux
Klan, the National Alliance, and a number of self-proclaimed Christian-right
organizations use their Web sites to broadly convey their message and attract
potential new members to their causes, often combining racist attacks with attacks
based on religion, gender, and sexual identity.9 As is often the case with hate
propaganda, the vitriol of many of the messages is camouflaged with attractive
packaging and, in some cases, pseudo-academic approaches.”0 Nevertheless, the
hateful nature of the messages is rarely far from the surface1’ and can be combined
information on bomb making or advocacy of other violent activity. 2
with
Connections between on-line hate propaganda and “real world” violence serve as a
individuals identifiable on the basis of personal characteristics such as race, religion, ethnicity, gender,
family status, marital status, and sexual identity that have historically formed the basis of socially
imposed disadvantage.
7 Anti-Defamation League, “Hate on the World Wide Web: A Brief Guide to Cyberspace Bigotry”,
online: Anti-Defamation League
Defamation League, “Hate on the Web”].
8 Wendy McAuliffe, “Europe Hopes to Outlaw Hate Speech Online” CNET News.com (12
November 2001), online: CNET
9 Southern Poverty Law Center, “Hate on the Internet,” online: Tolerance.org
10 Two Web sites recently subjected to Canadian Human Rights Tribunal (“CHRT’) cease and desist
orders, for example, attempt to accredit their messages of hate by associating them with “noted
historians” such as David Irving, or publishing organs such as the Institute for Historical Review, an
organization that claims to be committed to debunking the “myth” of the Holocaust. See Citron v.
Zundel, [2002] 41 C.H.R.R. D/274 T.D.1/02 (CHRT) [Citron] and Warman v. Kyburz, 2003 CHRT 18
[Warman].
11 For example, the home page of one anti-gay and lesbian Web site, recently ordered shut down by
a CHRT, states: “Warning! This site contains material that is deemed offensive by homosexual
pedophiles” (Citizens Research Institute, online:
12 For example, the Creativity Movement (formerly the World Church of the Creator) operated a
Web site known as Skinheads of the Racial Holy War, though the group disclaimed participation in or
advocacy of violence: Anti-Defamation League, “World Church of the Creator: ‘Racial Holy War’ on
the Web”, online: Anti-Defamation League
Web sites advocate or condone violence in a somewhat less direct fashion by posting articles
advocating extermination and violence. See e.g. Warman, supra note 10 at paras. 20-26.
20041
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
65
chilling reminder of the potential impact of words on action, particularly words
targeted at vulnerable consumers such as children. 3
It may well be that the reported scope of Internet hate propaganda represents only
the tip of the iceberg. Hate group activity has spread from Web sites to other Internet
applications, such as e-mail, chatrooms, and newsgroups, which are currently more
difficult to track. Further, the Internet is used to advertise and distribute other hateful
materials, such as music and video games. 4 By restricting access
to these
applications, hate groups can foment and reinforce hateful attitudes and approaches
among like-minded individuals while avoiding the backlash of negative publicity
created when their messages are readily available to other Internet users. In other
cases, these forums are generally available to all members of the connected public. 5
These applications may be used to incite “lone-wolf acts” as hate groups attempt to
distance their organizations from violent action resulting from their messages. 6
The known scope of Internet hate propaganda has prompted national and
international attention and calls for action, focusing predominantly on racist
propaganda. The Council of Europe (“COE”) held an international forum in 2001
relating to illegal and harmful content on the Internet, in which racist propaganda was
a central focus. 7 Subsequently, a number of COE member states signed a protocol to
address racist and xenophobic acts through computer networks (the “Additional
Protocol”). 8 Although Canada has yet to sign the Additional Protocol, it has taken
action at the national level, amending provisions of both the Canadian Human Rights
Act 9 and the Criminal Code (“Code”) to specifically address dissemination of
the “Amendments”). 21 The
hatred
through computer networks
(collectively,
” See Peter J. Breckheimer II, “A Haven for Hate: The Foreign and Domestic Implications of
Protecting Internet Hate Speech Under the First Amendment” (2002) 75 S. Cal. L. Rev. 1493 at 1496-99.
14 See Anti-Defamation League, “Hate on the Web”, supra note 7; Franklin, supra note 9.
15 The CHRT in Warman (supra note 10 at paras. 57-60) found that the hate propagandizing tactics
of the respondent in that case included use of group e-mail and a Web forum in which membership
was open generally to the connected public.
16 See Anti-Defamation League, “‘Lone Wolf of Hate Prowls the Internet” (2000), online: Anti-
Defamation League
17 Council of Europe, “European Forum on Harmful and Illegal Cyber Content” (22 November
2001), online: Council of Europe
Telecommunications Union, World Summit on the Information Society, “Draft Plan of Action” (12
December 2003) DOC.WSIS-O3/GENEVA/Doc-5-E, online: International Telecommunications Union
18 Additional Protocol to the Convention of Cybercrime, Concerning the Criminalisation ofActs of a
Racist and Xenophobic Nature Committed Through Computer Systems, 28 January 2003 Eur. T.S. 189
[Additional Protocol]. For the Chart of Signatures and ratifications to the Additional Protocol, see
online: Council of Europe
[Chart of
signatures].
‘9 Canadian Human Rights Act, R.S.C. 1985, c. H-6 [CHRA].
20 Criminal Code, R.S.C. 1985, c. C-46 [Code].
21Anti-Terrorism Act, S.C. 2001, c. 41, ss. 10, 88 [ATA].
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
Amendments are consistent with the Canadian constitutional approach to restrictions
on hate propaganda, as well as Canada’s international human rights obligations.
However, the First Amendment continues to protect a significant body of Internet
hate propaganda emanating from servers located in the United States.
II. Canadian Approach to Hate Propaganda
A. Amendment of the CHRA and the Code
The federal government’s first direct legislative attempt to explicitly restrict
Internet hate propaganda arose in the context of the Anti-Terrorism Act-legislation
designed to address terrorism following the attacks on the US on 11 September
2001.2 The ATA extends to the Internet specific pre-existing Criminal Code23 and
CHRA24 restrictions on defined categories of hate propaganda, which were previously
determined by the Supreme Court of Canada to be consistent with freedom of
expression under the Charter (as discussed in detail below in subsection B).
The objectives of the ATA and the context surrounding its passage underscore
continuing concern about links between hate propaganda, intolerance, and violence.
Then Justice Minister Anne McClellan specifically linked the objectives of the ATA
to the context of the September 11 attacks, stating:
The horrific events of September 11 remind us that we must continue to work
with other nations to confront terrorism and ensure the full force of Canadian
law is brought to bear against those who support, plan and carry out acts of
terror-we will cut off their money, find them and punish them.25
Representatives of the federal government characterized the Amendments as a
mechanism for addressing the root causes of hatred underlying terrorist acts, 26 stating:
22 Ibid.
23 Code, supra note 20.
24 CHRA, supra note 19.
25 Department of Justice Canada, News Release, “Government of Canada Introduces Anti-Terrorism
Act” (15 October 2001), online: Department of Justice Canada
26 In the House of Commons, Liberal members of Parliament relied on the Amendments in
responding to harsh criticisms by members of the opposition parties relating to other provisions of the
ATA that create a significant risk of targeted law enforcement against, and surveillance of, individuals
from certain religious and cultural backgrounds. See e.g. House of Commons Debates, 095 (16
October 2001) at 1705 (Libby Davies) and 1230 (Sarmite Bulte), online: Government of Canada
[“Bulte Debate”];
House of Commons Debates, 095 (17 October 2001) at 1625 (Anita Neville), online: Government of
Canada
Roach, eds., The Security of Freedom: Essays on Canada ‘ Anti-Terrorism Bill (Toronto: University of
Toronto Press, 2001).
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
67
This is a struggle against terrorism, and not against any one community, group
or faith. Diversity is one of Canada’s greatest strengths, and the Government of
Canada is taking steps to protect it. Measures will be included … to address the
root causes of hatred and to ensure Canadian values of equality, tolerance and
fairness are affirmed in the wake of the September 11 attacks. These include:
*
”
amending
propaganda … ; and
the Criminal Code to eliminate online hate
amending the Canadian Human Rights Act to clarify that the
prohibition against spreading
repeated hate messages by
telephonic communications includes all telecommunications
technologies.
27
The federal government linked the Amendments to the “war on terrorism” in two
ways. First, restricting messages inciting racial hatred might reduce the risk of
incitement of terrorist conduct. Second, such restrictions should protect those of
certain identifiable racial, ethnic, and religious origins from being wrongfully
subjected to hatred or contempt as the result of terrorist acts.2 s
The Amendments comprise two components-one relating to the CHRA (the
“CHRA Amendment”) and the other to the Criminal Code (the “Code Amendment”).
1. CHRA Amendment
Section 88 of the ATA amended subsection 13(2) of the CHRA by specifically
indicating that provisions of 13(1) apply to computer and Internet communication.29
As amended, subsections 13(1) and 13(2) provide that it is a discriminatory practice
for a person or a group of persons acting in concert to use, among other things, a
computer or group of interconnected or related computers to communicate or to cause
to be communicated, repeatedly, any matter that is likely to expose a person or
persons to hatred or contempt by reason of their identifiability based on a prohibited
ground of discrimination. The prohibited grounds of discrimination under the CHRA
include race, national or ethnic origin, colour, religion, age, sex, sexual orientation,
marital status, family status, disability, and conviction for which a pardon has been
granted.30
Subsection 13(1) may be applied
those communicating or causing
communication of messages located outside of Canada, provided that the victim of
the practice was, at the time, a citizen or permanent resident of Canada.3′ In the
Internet context, this may also include certain Internet Service Providers (“ISP”s)
to
27 Department of Justice Canada, supra note 25.
28 See e.g. “Bulte Debate”, supra note 26; “Neville Debate”, supra note 26.
29 In Citron, supra note 10, a CHRT had already determined prior to passage of the clarification in
the CHRA Amendment that subsection 13(1) applied to Intemet communication via the World Wide
Web.
30 CHRA, supra note 19, s. 3(1).
31 Ibid., s. 40(5)(c).
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
involved in causing communication, unless their only role in the communication is by
virtue of others using their undertaking for purposes of transmission.32
Where a discriminatory practice is made out under subsection 13(1), a Canadian
Human Rights Tribunal (“CHRT”) may require the person found to have engaged in
such a practice to: (i) cease that practice and to take measures in consultation with the
Canadian Human Rights Commission (“CHRC”) to redress the practice or prevent it
from recurring; (ii) compensate a victim specifically identified in the communication;
and (iii) pay a penalty of up to $10,000. 3
2. Code Provisions and Amendment
reflecting Canada’s
intemational human
rights
Even prior
to
the ATA,
obligations,34 it was a crime in Canada to:
(i) advocate killing or deliberately inflicting conditions of life calculated to
bring about the physical destruction of a section of the public identifiable on
the basis of colour, race, religion, or ethnic origin (an “Identifiable
Group”);
35
(ii) publicly incite hatred against an Identifiable Group where the incitement was
likely to lead to a breach of the peace; 36 and to
(iii) publicly communicate statements willfully promoting hatred against an
Identifiable Group (subject to the defences of good faith and truth, among
others).37
Section 10 of the ATA amended the Code to explicitly empower a judge to
address on-line hate propaganda by: (i) issuing a warrant of seizure of, among other
things, hate propaganda as defined in sections 318 and 319 of the Code (“Hate
Propaganda”) stored on and made publicly available through a computer system
32Ibid., s. 13(3).
Ibid., s. 54(1).
34 See especially International Convention on the Elimination of All Forms of Racial
Discrimination, 4 January 1969, 660 U.N.T.S. 212 [CERD] (signed by Canada 24 August 1966 and
ratified 14 October 1970); Canada, Law Reform Commission, Hate Propaganda (Ottawa: The
Commission, 1986) at 17; International Covenant on Civil and Political Rights, 23 March 1976, 999
U.N.T.S. 172 [ICCPR].
35 Code, supra note 20, s. 318. The prohibited grounds of discrimination under the Code are notably
more restricted than those under the CHRA. However, a majority of the House of Commons voted on
17 September 2003 to include sexual orientation as a prohibited ground in the Code provision. See
Bill C-250, An Act to amend the Criminal Code (hate propaganda), 2d Sess., 37th Parl., 2002 and
Peter O’Neil, “Homosexuals to be Covered by Anti-Hate Legislation” Ottawa Citizen (18 September
2003) A5.
36 Code, ibid., s. 319(1).
3 7 Ibid., s.319(2).
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
69
within the jurisdiction of the court” and, after a hearing, (ii) ordering the seized
material deleted if it meets the statutory criteria.39 While the hearing regarding
deletion may involve individuals residing outside of Canada,4 it is questionable
whether deletion orders will be made in respect of computer systems in other
jurisdictions.1
B. Canadian Constitutional Review of Hate Propaganda Restrictions
Prior to enactment of the Amendments, the Supreme Court had found, in Taylor4′
and Keegstra” respectively, that the restrictions imposed by subsection 13(1) of the
CHRA and by subsection 319(2) of the Code constituted justifiable limits on
expression in a free and democratic society. The Court concluded that these
provisions restricted non-violent attempts to convey meaning and thus violated
subsection 2(b) of the Charter,’ but were nevertheless justifiable in that:
(i) Hate propaganda as defined in the provisions lay far from the core values of
the search for the truth,45 democratic participation,” and self-fulfillment”
38Ibid., s. 320.1(1).
39 Ibid., s. 320.1(5).
40 Ibid. Subsection 320.1(2) requires notification of the person posting the offending information
with respect to the impending hearing as well as the opportunity to attend and to make submissions
with respect to why the material should not be ordered deleted. See also subsection 320.1(4), which
specifically contemplates that the person posting the material may reside outside of Canada and
provides that a deletion order may be made even if the person posting the material does not attend the
hearing.
41 While a court could theoretically interpret “within the jurisdiction of the court” (ibid., s. 320.1(1))
widely and attempt to assert control over a computer system located outside of Canada, the legislative
intention behind the provision appears to require that the computer system must be physically present
in Canada in order to issue a warrant in relation to it. See House of Commons Debates, 095 (16
October 2001) at 1015 (Anne McLellan), online: Government of Canada,
practice, if investigative authorities had reasonable grounds to believe that evidence in the US would
assist in proving a violation within Canada of the hate propaganda provisions under the Criminal
Code through a computer system, it is likely that Canadian authorities would request their American
counterparts to obtain a US search warrant pursuant to the Treaty Between the Government of Canada
and the Government of the United States ofAmerica on Mutual Legal Assistance in Criminal Matters,
Canada and the United States, 18 March 1985, Can. T.S. 1990 No. 19 (entered into force 24 January
1990). Under the treaty, the US Central Authority may deny assistance to the extent it considers
Canada’s request to be contrary to US public policy (art. (V)). Further, the US Central Authority
would be required to obtain a warrant, at which point a US court’s willingness to issue could be
affected by whether the warrant sought related to First Amendment protected expression.
42 Canadian (Human Rights Commission) v. Taylor, [1990] 3 S.C.R 892, 75 D.L.R. (4th) 577
[Taylor cited to S.C.R.].
43 R. v. Keegstra, [1990] 3 S.C.R. 697, [1991] 2 WW.R. 1 [Keegstra cited to S.C.R.].
44 See Irwin Toy Ltd. v. Quebec (A.G.), [1989] 1 S.C.R. 927, 58 D.L.R. (4th) 577.
45 Applying this approach, the Court found that the restricted expression did not facilitate the search
for truth, noting both that it was premised upon untruths and partial truths, and that it could not be
assumed that an unregulated marketplace of ideas would necessarily facilitate the search for truth
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
underlying freedom of expression, making their restriction more easily
justifiable; a”
(ii) The Code and CHRA provisions served pressing and substantial objectives
underscored by other Charter values such as equality and multiculturalism,
as well as Canada’s international human rights obligations,49 respectively
being aimed at: limiting the risk of harm that hate propaganda poses to target
group members and to racial, ethnic, and religious harmony in Canada” and
promoting equality of opportunity unhindered by discriminatory practices
based upon membership in, among others, a particular racial, religious, or
ethnic group;
(iii) Prohibiting
the
provisions was rationally connected with their objectives in that censure of
the dissemination of hate propaganda as defined
in
(Keegstra, supra note 43 at 747). The dissenting reasons in Zundel echoed these conclusions, noting
that “[w]e are warned quite properly that history has many lessons to teach. One is that the
marketplace of ideas is an inadequate model; another is that minorities are vulnerable to censure as
speakers”: R. v. Zundel, [1992] 2 S.C.R. 731 at 825, 95 D.L.R. (4th) 202 [Zundel cited to S.C.R.].
While the majority in Zundel held that untruths could play a useful role in the search for truth, it is
unlikely that the majority view in that case would bear upon an analysis of the constitutionality of the
Amendments. The Court’s more recent comments in Thomson reaffirm that harmful effects associated
with hate propaganda render restrictions on it more easily justified: Thomson Newspapers Co. v.
Canada (A.G.), [1998] 1 S.C.R. 877 at paras. 90-92, 159 D.L.R. (4th) 385 [Thomson cited to S.C.R.].
Further, the provision at issue in Zundel related to “false news”-a restriction originally aimed at
protecting the aristocracy from scandalous remarks. The Amendments, are aimed at, among other
things, promoting social harmony.
46 The Court found that hate propaganda argues “for a society in which the democratic process is
subverted and individuals are denied respect and dignity simply because of racial or religious
characteristics … [a] brand of expressive activity … wholly inimical to the democratic aspirations of
the free expression guarantee” (Keegstra, supra note 43 at 764).
47 The Court accepted that the hate propaganda at issue facilitated the individual self-fulfillment of
the speaker, but undennined the fulfillment of target group members, reasoning that
self-autonomy stems in large part from one’s ability to articulate and nurture an identity
derived from membership in a cultural or religious group. The message put forth by
individuals who fall within the ambit [of the hate provisions of the Code] represents a
most extreme opposition to the idea that members of identifiable groups should enjoy
this aspect of the 2(b) benefit. The extent to which the unhindered promotion of this
message furthers free expression values must therefore be tempered insofar as it
advocates with inordinate vitriol an intolerance and prejudice which views as execrable
the process of individual self-development and human flourishing among all members
of society (ibid. at 763).
48 See Keegstra, ibid., as reaffinmed in Thomson, supra note 45 at paras. 90-92.
49 See Keegstra, ibid. at 746-50.
50 See ibid. at 758.
51 See Taylor, supra note 42 at 918.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
71
the expression restricted fostered the protection of target group members and
promoted equality, diversity, and multiculturalism in Canadian society; 2 and
(iv) The provisions were
tailored
than private
dissemination of the expression in issue and, as such, their likely salutary
effects on equality, multiculturalism, and protection of target group members
outweighed their deleterious impact on expression.53
to restrict public,
rather
The Court concluded that the provisions in question were consistent with
Canada’s international human rights obligations under the International Convention
on the Elimination of All Forms of Racial Discrimination54 and the International
Covenant on Civil and Political Rights,55 which oblige signatory states to respect
freedom of expression and prohibit hate propaganda, suggesting that the two are not
necessarily mutually exclusive. The Court further noted that a Human Rights
Committee, appointed under the Optional Protocol to the ICCPR, dismissed Taylor’s
complaint that subsection 13(1) of the CHRA violated his right to free expression.
The Committee held that “the opinions which Mr. Taylor seeks to disseminate
through the telephone system clearly constitute the advocacy of racial or religious
hatred which Canada has an obligation under article 20(2) of the [ICCPR] to
prohibit.”56
Although decided without regard for the Amendments, two recent CHRT
decisions under the pre-amendment version of subsection 13(1) of the CHRA relied
on the Taylor and Keegstra constitutional analyses to conclude that the provision
remains valid, even as applied to hate propagation through World Wide Web
52 The Court found that the provisions were rationally connected to these objectives in three ways.
First, limiting or eliminating hate propaganda reduced the risk that target group members would be
harmed through direct exposure to it. Second, the restrictions on dissemination of hate propaganda
reduced
the risk that equality, diversity, and multiculturalism would be undermined through
recruitment of Canadians to the cause of racism. Third, the restrictions served to publicly denounce
hate propaganda, thereby affirming the Canadian constitutional values of multiculturalism and
equality (see Keegstra, supra note 43 at 769-71; Taylor, ibid at 922-24).
53 In making this determination, the Court emphasized the constitutional imperative of protecting
private spheres of communication from undue govemment intrusion. It noted that the Code provisions
apply only to comments willfully made publicly accessible (see Keegstra, ibid. at 771-79), and that
the CHRA provision would likely apply only to public forms of communication given the threshold of
“hatred” imposed in the provision (see Taylor, ibid. at 936-38). The latter conclusion is buttressed by
the Court’s later decision that human rights code provisions are unlikely to apply to private
organizations of intimates. See Gould v. Yukon Order of Pioneers, [1996] 1 S.C.R. 571, 133 D.L.R.
(4th) 449.
54 CERD, supra note 34, art. 4.
55 ICCPR, supra note 34, art. 20(2).
56 Taylor and Western Guard Party v. Canada, Communication No. 104/1981, Report of the Human
Rights Committee, 38 U.N. GAOR, Supp. No. 40 (A/38/40) 231 (1983), decision reported in part at
(1983), 5 C.H.R.R. D/2097.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
(“WWW”) sites.57 This approach was more recently cited with approval by the CHRT
in Warman, a case in which there was some question whether the pre-amendment or
post-amendment version of subsection 13(1) applied. The prior constitutional
conclusions reached in Taylor and Keegstra continue to apply, perhaps with even
to restrictions on hate propaganda expressed via the Internet. 9
greater force,
However, despite the constitutionality of the Amendments and pre-existing Code
restrictions on hate propaganda in Canada (collectively, the “Canadian Provisions” or
“Provisions”) and their consistency with Canada’s
international human rights
obligations,
them will be challenged by First
Amendment protection against direct restriction of hate propaganda emanating from
US servers.
the public policy underlying
III. First Amendment Protection of Hate Propaganda
The First Amendment states “Congress shall make no law abridging freedom of
speech, or of the press,” and generally protects hate speech.6″ Under the First
Amendment, orders such as those likely to be sought or issued under legislation like
the Canadian Provisions would be analyzed in two steps to determine: (i) whether
they are directed at the communicative impact of expressive activity, and if so; (ii)
whether the expression in question is excluded from First Amendment protection.61
Most orders issued under laws like the Canadian Provisions are likely to be aimed
at minimizing the communicative impact of Internet hate propaganda, and thus to be
considered a prima facie violation of the First Amendment.62 Their constitutionality
would then depend on demonstrating that they are aimed at unprotected expression-
57 Citron, supra note 10; Schnell v. Machiavelli and Associates Emprize Inc. and John Micka, T.D.
11/02 (CHRT). For a more detailed discussion of these decisions, see Jane Bailey, “Of Mediums and
Metaphors: How a Layered Methodology Might Contribute to Constitutional Analysis of Internet
Content Regulation” Man. L.J. [forthcoming in 2004].
58 Warman, supra note 10 at paras. 12-15, 55.
59 Additional justification for restricting hate propaganda can be found in the Internet’s tremendous
potential breadth of dissemination, as well as a trend among Internet hate speakers to use key words
that might cause unintentional accessing of their sites by unsuspecting Web searchers, presumably in
order to attract attention to their “causes”. See Breckheimer, supra note 13 at 1498-99.
60 For a more detailed analysis of the evolution of US law relating to hate speech, including the
constitutionality of legislation addressing racist expression less directly than the Canadian Provisions
(such as increased penalties for racially motivated crimes), see Rachel Weintraub-Reiter, “Hate
Speech Over the Internet: A Traditional Analysis or a New Cyber Constitution” (1998) 8 B.U. Pub.
Int. L.J. 145 and William B. Fisch, “Hate Speech in the Constitutional Law of the United States”
(2002) 50 Am. J. Comp. L. 463.
61 See Laurence Tribe, American Constitutional Law, 2d ed. (Mineola, N.Y.: The Foundation Press,
1988) at 791-94.
62 See ibid.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
73
expression falling within an excluded category or the limitation of which is underlain
by a “compelling state interest” 3
Hate propaganda might have been interpreted to fall within one of the excluded
categories identified in early US jurisprudence, such as words of “slight social value”,
libel, “fighting words”, or words creating a clear and present danger of imminent
violence.’ Although
the US Supreme Court in one instance found that hate
propaganda fell within the libel exclusion,65 subsequent decisions by US courts, such
as that of the US Supreme Court in R.A. V,66 suggest that the excluded categories will
be very narrowly interpreted and explicit restrictions on hate propaganda strictly
scrutinized under the First Amendment.67 Thus, enforceability is likely to depend
upon demonstrating a compelling state interest.6″
It is likely to be difficult for orders under legislation like the Provisions to meet
the “compelling state interest” standard currently applied in the United States. US
courts have found that direct prohibitions on parading in Nazi dress in a town in
which the majority of residents are Jews69 and on burning a cross on the lawn of a
racialized family do not address sufficiently compelling state interests to pass
constitutional muster.7″ Further, subsequent decisions suggest restrictions on Internet
in general will be strictly scrutinized.7′ Nevertheless, some
communication
63 See ibid.
64 See Schenckv. United States, 249 U.S. 47 at 52 (1919), 39 S. Ct. 247. See also Chaplinsky v. New
Hampshire, 315 U.S. 568 at 572 (1941).
65 See Beauharnais v. Illinois, 343 U.S. 250 (1952), 72 S. Ct. 725.
66 R.A. V v. City of St. Paul, 505 U.S. 377 at 2542-44 (S. Ct. 1992), 112 S. Ct. 2538 [R.A. V cited to
U.S.] (finding that the government may not directly prohibit acts that are known or should be known
to arouse anger, alarm, or resentment on the basis of race, colour, creed, religion, or gender).
67 See Village ofSkokie v. National Socialist Party of America, 366 N.E.2d 347 (Ill. App. Ct. 1977)
51 111. App. 3d 279, rev’d in part, 373 N.E.2d 21 (S. Ct. Ill. 1978), 69 Ill.2d 605 [Village of Skokie
cited to N.E.2d] (affirming the right to parade in Nazi dress, displaying swastikas in a town in which
the majority of residents were Jews) and Black v. Commonwealth, 262 Va. 764, 553 S.E.2d 738
(following Village of Skokie).
68 See Tribe, supra note 61 at 833-34. Orders issued under laws like the Canadian Provisions will
only meet the “compelling state interest” test if it is shown that: (i) the state interest was actually
considered by the legislator, (ii) a close nexus exists between the means chosen and the state interest
to be served, and (iii) the limitation on expression is narrowly drawn. Satisfaction of the second and
third criteria will depend largely upon the drafting of any particular order and is therefore more
usefully assessed in the context of specific orders. In any event, the latter two criteria will only be
relevant if the compelling state interest criterion can be satisfied.
69 See R.A. V., supra note 66.
70 See Village of Skokie, supra note 67.
71 See Yahoo! Inc. v. La Ligue Contre Le Racisme et L’Antisemitisme, 169 F.Supp.2d 1181 at 1186
(N.D.Cal. 2001) [Yahoo! v. Ligue] (declaring unenforceable a French court order requiring a US-based
ISP to prevent French citizens from accessing on-line auctions of Nazi memorabilia) and Reno v.
American Civil Liberties Union, 521 U.S. 844, 117 S. Ct. 2329 (1997) [Reno cited to U.S.] (virtually
equating the Internet with the ideal marketplace of ideas, requiring strict scrutiny of restrictions on
Internet communication).
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
restrictions might be found to address the excluded category of expression presenting
a clear and present danger of imminent harm, such as hate propaganda taking the
form of “true threats” of violence made against a particular individual72 or of broadly
based terrorist threats.73 The compelling nature of the state interest in regulating the
second type of hate propaganda is arguably made more apparent by the September 11
attacks.7 4 In either situation, however, an order is likely to be enforceable only where
it is clear that the hate propaganda at issue is likely to lead to immediate action
without the opportunity to avoid harm through response and discussion.75
IV. Conflicting Democratic Visions
The constitutionality of the Canadian Provisions in Canada and the likely
unconstitutionality in the US of most orders issued under like provisions demonstrate
the fundamentally different approaches to hate propaganda taken by two countries
sharing a common commitment to democracy. The constitutional validity of the
Provisions in Canada reflects other fundamental democratic values in addition to free
expression, including multiculturalism and equality, which are also protected by
international human rights instruments such as the CERD and ICCPR. Although the
US is also a signatory to these instruments,76 most orders issued pursuant to
72 See Weintraub-Reiter, supra note 60 at 147-48, citing US. v. Alkhabaz, 104 F.3d. 1492 (6th Cir.
1997) [Alkhabaz].
73 See Cass Sunstein, “Constitutional Caution” (1996) U. Chicago Legal F. 361 at 366-72.
74 Like Canada, the US has enacted legislation in response to the 11 September 2001 attacks. The
USA PATRIOTAct No. 107-56 (26 October 2001) significantly expands surveillance and investigative
powers with respect to “terrorism”. While the Senate preamble of the act expresses concern for and
condemns violent acts toward Arab, Muslim, and South Asian Americans and focuses on electronic
communications, none of its provisions parallel the Amendments in terms of restricting hate
propaganda. For an analysis of the act, see Electronic Frontier Foundation, “EFF Analysis of the
Provisions of the USA PATRIOTAct that Relate to Online Activities” Electronic Frontier Foundation
(31 October 2001), online: Electronic Frontier Foundation
legislative signal of national security
concerns could prompt a different judicial approach to restrictions on hate propaganda that could be
brought within the rubric of terrorism.
75 See Tribe, supra note 61 at 791, 833-34. Moreover, the applicability of this narrow exception is
debatable in the Internet context. Some recent US decisions and academic commentary suggest that
the physical distance that often separates the speaker from the listener in Internet communications
may make it difficult to establish that harm is sufficiently imminent to justify restricting speech:
Alkhabaz, supra note 72; Stuart Biegel, Beyond Our Control? Confronting the Limits of Our Legal
System in the Age of Cyberspace (Cambridge, Mass.: M.I.T. Press, 2001) at 339-40.
76 Although the US ratified the ICCPR on 8 September 1992 and the CERD on 20 November 1994,
it placed reservations on their ratifications, confirming that its obligations under these treaties will be
interpreted in accordance with freedom of expression under the US. Constitution. See United Nations
High Commissioner of Human Rights, Status of Ratifications of the Principal International Human
Rights Treaties (7 July 2003), online: United Nations, Office of the High Commissioner for Human
Rights
Reservations (5 February 2002), online: United Nations, Office of the High Commissioner for Human
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
75
legislation like the Canadian Provisions (however tailored and specific) are likely to
be considered unenforceable violations of the First Amendment.
The difference in these two approaches does not reflect happenstance. The Court
in Keegstra, adverted to the US approach to hate propaganda and deliberately chose
to follow a different path. The majority reasoned that:
Where s. 1 operates to accentuate a uniquely Canadian vision of a free and
democratic society, however, we must not hesitate to depart from the path taken
in the United States. Far from requiring a less solicitous protection of Charter
rights and freedoms, such independence of vision protects these rights and
freedoms in a different way. … [1]n my view the intemational commitment to
eradicate hate propaganda and, most importantly, the special role given equality
and multiculturalism in the Canadian Constitution necessitate a departure from
the view, reasonably prevalent in America at present, that the suppression of
hate propaganda is incompatible with the guarantee of free expression. 77
The path chosen by Canadian legislators and approved by the Court is one no less
informed by a fundamental commitment to democracy than that adopted by the US
Supreme Court. In fact,
is by other
fundamental and internationally recognized democratic values, such as equality and
multiculturalism, arguably reflects a more comprehensive conception of both private
and public forces affecting individual liberty than that adopted in the US. As I have
suggested elsewhere, 8 materially different determinations as to the proper role of the
state underlie the different approaches taken in Canada and the US.
the Canadian approach,
informed as it
The Canadian approach recognizes both public and private sources of oppression
on individual liberty and expressly, through section 1 of the Charter, acknowledges
the potential role for government in ameliorating the negative impacts of private
sources of oppression on individuals.79 In contrast, the current US approach” focuses
almost exclusively on the state as the source of oppression and reflects an inherent
distrust of state authority, even when used to address conflicts between groups of
individuals.8′
The cross-border flow of First Amendment protected hate propaganda from the
US, facilitated by the Internet, poses a threat to the broader goals of multiculturalism
Rights
unhchr.ch/html/menu3/b/treatyl9asp.htm> (for the ICCPR).
77 Keegstra, supra note 43 at 743.
78 Bailey, supra note 57.
79 Dickson C.J.C. (as he then was) cautioned that “one must be careful not to accept blindly that the
suppression of expression [by government] must always and unremittingly detract from values central
to freedom of expression” (Keegstra, supra note 43 at 765).
(for the CERD) and
working group was established in April 2002 to provide further working drafts. The first of these
relates to the choice of court in commercial matters. See Hague Conference on Private International
Law, Permanent Bureau, “Preliminary Result of the Work of the Informal Working Group on the
Judgments Project” Preliminary Document No. 8 of March 2003, online: Hague Conference on
Private International Law
draft, see Andrea Schulz, “Report on the Work of the Informal Working Group on the Judgments
Project, in Particular on the Preliminary Text Achieved at Its Third Meeting 25-28 March 2003” (June
2003),
87Draft Hague Convention, ibid., art. 1.
International
online:
on
Private
Law
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
Although not applicable to orders issued under criminal or administrative
legislation, such as the Canadian Provisions, models such as the Draft Hague
Convention could be used for inter-jurisdictional enforcement of restrictions on
Internet hate propaganda in a number of ways. They might be used to enforce a civil
award for posting material libeling a racially identifiable group in one jurisdiction
against a speaker in another jurisdiction. They could also be used to enforce a damage
award issued in one jurisdiction for failing to provide an on-line environment free
from racial discrimination against an ISP s in another jurisdiction.
However, it is unlikely that these procedural mechanisms would be significantly
more effective than the Canadian Provisions alone in restricting cross-border flows of
hate from the US. Signatory states are certain to be explicitly entitled to refuse to
enforce or recognize foreign orders they consider “manifestly incompatible with
[their] public policy.” 9 The US could be expected to rely on a provision such as this
to refuse to enforce orders restricting hate propaganda.9″
B.
International Agreements Harmonizing Substantive Law: The
Additional Protocol to the Cybercrime Convention
In 2001,
thirty-three nations,
including Canada and the US, signed
the
Cybercrime Convention, which provides, among other things, for the harmonization
of laws among signatory states with respect
to computer crimes and child
pornography on the Internet.9 While a number of European nations pressed for
inclusion of a provision relating to hate propaganda during initial negotiations, the
US successfully resisted, citing First Amendment concerns.92 Recent experience
regarding the Additional Protocol suggests the improbability of a state-based
88 As discussed in greater detail, infra, a customer unsuccessfully sued AOL on the basis of this
cause of action in the US. For the pleading, see Statement of Claim for Noah v. AOL lime Warner
(August 2001), online: Council for American-Islamic Relations
F Supp. 2d 532 (E.D. Va. 2003) [Noah].
89 Draft Hague Convention, supra note 86, art. 28(1)(f).
90 As it has done with respect to the ICCPR, supra note 34, and the CERD, supra note 34. Similar
concerns arise with respect to proposed common law tests for determining whether a court should
accept jurisdiction in any given case. With respect to “targeting”, see Michael Geist, “Is There a There
There? Toward Greater Certainty for Internet Jurisdiction” (2002) 16 Berkeley Tech. L.J. 1. Even if a
US court were to accept that a Canadian court properly asserted jurisdiction over a Web site targeted
at Canadians, public policy considerations would likely prevent enforcement of the Canadian
judgment in the US. Furthermore, in the hate propaganda context, where the speaker’s objective often
appears to be to target anyone who will listen, it may be difficult to identify a “target” to establish
jurisdiction.
91 Cybercrime Convention, 23 November 2001, Eur. T.S. No. 185, art. 9, online: Council of Europe
92 Carl S. Kaplan, “New Economy: Bracing for a Flood of Efforts to Control Speech Seen as
Hateful or Terrorist” The New York Times (11 February 2002) C3, online: The New York Times
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
79
international response to potential First Amendment hegemony over Internet hate
propaganda.93
The COE approved the Additional Protocol in January, 2003. 94 It requires, subject
to certain reservations, that signatory states establish criminal offences for using a
computer system to:
a. Make racist and xenophobic material publicly available;
b.
threaten a serious criminal offence against or publicly insult a person or
persons due to their membership in a group distinguished by race, colour,
descent or national or ethnic origin;
deny, grossly minimize, approve of, or justify acts of genocide or crimes
against humanity; and
aid and abet commission of any of these offences.95
c.
d.
It defines “racist and xenophobic material” to include written material, images, and
other representations of ideas or theories advocating, promoting, or inciting hatred,
discrimination, or violence against an individual or group based on race, colour,
descent, or national or ethnic origin, as well as on religion to the extent it is used as a
“pretext” for the other factors.96
By June, 2003, sixteen COE members had signed the Additional Protocol,
although none of the non-member states (including Canada) had done so. 97 Even
before its approval by the COE, the Bush administration announced that the US
would not support the Protocol, again citing First Amendment concerns.98 At least in
the short-term, there is likely to be little remaining political will on the part of the US
and COE member states to negotiate any form of international agreement calling for
state harmonization of substantive law with respect to Internet hate propaganda. In
any event, so long as international agreements, such as the Additional Protocol,
permit each signatory state to restrict compliance with its obligations according to its
own public policy, this kind of arrangement is unlikely to expand the scope of legally
93 US refusal to sign the Additional Protocol, however, does not necessarily indicate that there is no
opportunity to harmonize national laws on this issue. The protocol would require legal restriction of
broad categories of expression that are not restrictable under the current First Amendment approach.
That is not to say, however, that there can be no co-operation, at least with respect to the narrow
category of hate propaganda not subject to First Amendment protection. The US Supreme Court’s
decision in U.S. v. American Library Association, Inc. 123 S. Ct. 2297, 156 L. Ed. 2d 221 [American
Library cited to S. Ct.], may also signal the possibility of restricting hate propaganda indirectly by
making requirements to limit its dissemination a condition of receiving federal funding or government
contracting opportunities.
94 Additional Protocol, supra note 18.
9′ Ibid., arts. 4-7. The Additional Protocol specifically permits signatories to make reservations to
avoid criminalizing the acts referred to in paras. (a) to (c), except for racially motivated threats.
96 Ibid., art. 2.
97 See Chart of signatures, supra note 18.
98 See Declan McCullagh, “U.S. Won’t Support Net ‘Hate Speech’ Ban” CNET News.com (15
November 2002), online: CNET
MCGILL LAw JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
restrictions against hate propaganda emanating
enforceable
the US.
Nevertheless, the Additional Protocol serves as an important first step in addressing
on-line hate propaganda and, to the extent it is widely ratified, may have persuasive
value in the development of US law and policy on this issue.
from
While the First Amendment limits the prospect for direct international public
regulation, private actors are not generally confined by constitutional restraints. Some
private actors in the US have in fact voluntarily undertaken to regulate on-line
expression, even in the absence of any legal compulsion to do so, opening up a
private and largely unscrutinized avenue for enforcement.
VI. Emerging Private Regulation
Private actors in the US and elsewhere are regulating on-line expression through
technological solutions (such as filtering and zoning), terms of use clauses in service
contracts, and self-regulatory organizations. While private regulation, generally free
from First Amendment compliance constraints, constitutes a poor substitute for public
regulation for both practical and policy reasons, the growing problem of Internet hate
propaganda suggests the need to consider both public and private strategies.
A. Privately Implemented Technological Solutions
The
two perhaps most frequently discussed
technological measures
for
addressing unwanted Internet content are filtering and zoning.99 Filtering enables
those at the receiving end of Internet communications to block content from being
delivered to them.” Zoning, which aims at a different link in the communication
chain, permits speakers to prevent certain users from accessing their content.10′ Critics
have questioned the technical ability of both measures to effectively disrupt the
delivery of particular content. Nonetheless, private implementation of filtering or
zoning at the user or ISP level has been the subject of a good deal of positive
commentary as a facilitator of individual choice.0 2
It will be suggested, however, that private implementation of filtering and zoning
fails to address the key social harms of concern in the context of hate propaganda: the
threat to social harmony and equality posed by widespread adoption of hate
propaganda’s message. Further, where implemented at the ISP level, filtering and
99 It is notable, however, that much of the discussion surrounding zoning and filtering focuses on
World Wide Web content. This need not be the case, as certain of these solutions can also be applied
to other Internet applications. For a relatively complete (albeit dated) discussion of this issue, see U.S.,
Commission on Online Child Protection, Final Report of the COPA Commission (20 October 2000),
online: Commission on Online Child Protection
100 See Lessig, supra note 1 at 175-76.
‘0’ See ibid.
102 See e.g. Weintraub-Reiter, supra note 60.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
81
zoning may be associated with the collection of private information with little or no
opportunity for public scrutiny.
1. Filtering
Individuals, organizations (such as ISPs3), or even countries (such as China) can
implement filtering, either to allow access only to materials identified as acceptable or
to block delivery of materials identified as unacceptable.” 4 In order to block
prohibited Internet content, someone, either the content provider or a third party, has
to rate it and apply labels to be read by users’ software applications. To the extent that
the user has chosen a software application configured for blocking certain content,
the application will deny access to that content based on the label attached to it.
Internet industry filtering initiatives in the early 1990s were largely designed to
address public concern about the ready availability of Internet content considered
harmful to minors.’0 5 The concept behind some of these efforts was to facilitate
individual user choice in accessing content, including choice not only in configuring
blocking preferences, but also in selecting rating systems compatible with personal
values and preferences.” The basic approach was to enable individual users to filter
out offensive content without interfering with the ability of others to access that
content should they so choose.
The current state of the art in filtering limits its efficacy. With the number of Web
pages alone growing exponentially, the sheer scope of the labelling project is
daunting to say the least,0 7 and technologies, such as encryption, permit users to
103 America Online, for example, relies on the software package produced by RuleSpace, a filtering
system capable of surveying the content of up to 47 million Web pages a day, to power the parental
control features that it markets as part of its service to subscribers: Jeffrey Benner, “AOL’s New Filter
on
the Block” Wired News (7 May 2001), online: Wired
104 Blocking might be carried out on a number of bases such as blocking reception of content
labelled hate propaganda, blocking all content emanating from a particular Uniform Resource Locator
(“URL”), or blocking key words associated with hate propaganda.
105 See World Wide Web Consortium (“W3C’), “Statement on the Intent and Use of PICS: Using
PICS Well” (1 June 1998), online: W3C
106 For example, non-governmental human rights groups, such as the Anti-Defamation League
(“ADL”), also manufacture filtering software, some of which is available free of charge on the
Internet. The ADL’s filtering program, known as HateFilter, allows consumers to filter Web content
based on the ADL’s philosophical approach and values. See Gwendolyn Mariano, “Anti-Hate Group
Updates Web Filter” CNETNews.com (21 March 2002), online: CNET
107 In the context of anti-social speech such as hate propaganda, leaving accurate labelling to
speakers is probably unrealistic. To the extent that the objective of hate propagandists is to reach as
wide an audience as possible, in the absence of public compulsion, they may have little incentive to
accurately label their content.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
circumvent filters imposed by others.’0 8 Further, the task of labelling either leaves out
critical aspects of human judgment (such as recognizing the importance of context) or
is done highly subjectively and with little opportunity for users to understand the
nuanced criteria employed by the labeler.’0 9 Given the exigencies of rating and
labelling, as well as circumvention techniques, expert evidence called at one US trial
suggested that, at that time, filters excluded innocuous material twenty-one per cent
of the time and offensive material only sixty-nine per cent of the time.”0 Figures such
as these may raise practical concerns about the current advisability of primary
reliance on privately
implemented filtering technology”‘ as a mechanism for
addressing Internet hate propaganda.” 2
More importantly, privately implemented filtering is inconsistent with at least one
material objective underlying the Canadian Provisions. It may be appropriate to
encourage private
individual
preferences in relation to legal content offending a particular user.”3 However,
Canadian restrictions on hate propaganda are not aimed solely at ensuring that
implementation of filters configured
to reflect
108 See Internet Law & Policy Forum, The Internet Law and Policy Forum Working Group on
Content Blocking (May 1997), online: ILPF
109 For a more thorough discussion of the criticisms of rating, see Jonathan Weinberg, “Rating the
Net” (1997) 19 Hastings Comm. & Ent. L.J. 453, online: Wayne State University Law School
110 Reuters, “Experts Lambaste Smut Filters” (26 March 2002) Wired News, online: Wired
Despite figures suggesting over-blocking, the US Supreme Court upheld legislation mandating public
libraries to filter content harmful to minors as a condition for receiving federal funding, noting that
adult patrons could request that the filters be disabled.
“‘1 On the other hand, if we were to examine the efficacy of legal restrictions alone, we might find
that technological solutions consistently delivering 69 per cent efficacy in eliminating access to
offensive material compare quite favourably with the results of legislative prohibitions.
112 See Lawrence Lessig & Paul Resnick, “Zoning Speech on the Internet: A Legal and Technical
Model” (1999) 98 Mich. L. Rev. 395. For these same practical reasons, the publicly compelled use of
filters, either by individual users, ISPs, or government is an unlikely policy alternative in the short
term. In addition, privacy concerns arise in the context of publicly mandated filtering, to the extent
that such measures require or encourage ISPs or governments to monitor the content accessed by
individual subscribers. Although international protocols obliging signatory nations to impose accurate
labelling requirements on their citizens could assist in the daunting task of content labelling, without
changes in the state of the art of the filtering software itself, questions remain as to both the efficacy of
this approach and the more pressing issues of the risks of monitoring and concerns about invasion of
privacy.
“3 For regulatory purposes, there is an important distinction between illegal content and legal, but
offensive, content. The European Council has conceptually segmented illegal from merely harmful
content, recognizing that the two merit different regulatory responses. See Beth Simone Noveck,
“European Forum on Harmful and Illegal Cyber Content: Rapporteur’s Report,” online: Council of
Europe
particularly meaningful one given that content is often determined to be illegal because it is harmful,
the attempt to delineate may be helpful in thinking about regulatory options.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
83
individual listeners are not offended or directly harmed. They also focus on reducing
or eliminating the risk of harm to the broader societal goals of equality and
multiculturalism posed by widespread adoption of hate propaganda’s message. ” 4
While filtering controlled by individual users or their ISPs may mitigate the
psychological harm suffered by target group members by reducing the risk that they
will be exposed to hate propaganda, it does not address these broader societal
concerns. Further, individually implemented filtering of hate propaganda makes it
invisible to those who might otherwise be mobilized to address it, while the hate
propaganda industry itself is permitted to thrive among those with attitudes that pose
the greatest risk of social harm.
2. Zoning
Internet speakers may prevent certain recipients from accessing their content
based on specific user characteristics. Technological developments
facilitate
recognition of personal characteristics of the user (such as age), as well as the user’s
geographic location (referred to as “geolocational” technology). Zoning based on the
user’s personal characteristics could, for example, assist a speaker in prohibiting a
child from accessing adult content posted on a Web site. Geolocational technology
assists speakers in denying access to users in countries where the content they seek to
access is illegal.115
Significant Internet industry players, such as AOL Time Warner Inc. (“AOL”)
and Google, rely on zoning software applications to build user profiles based on
hundreds, and even thousands, of demographic attributes of Web site visitors.” 6
Others, such as on-line gambling businesses, use geolocational technology to deny
access to those in jurisdictions where their content is prohibited or subject to severe
restriction. 117 Similarly, US-based hate speakers or ISPs providing Internet services to
hate speakers could implement geolocational technology to prevent citizens
in
countries where it is illegal from accessing that content.1 8 However, while some ISPs
not creating their own content may be so inclined, hate speakers are unlikely to
implement zoning technology. Businesses offering on-line gambling are likely to be
114 See Taylor, supra note 42; Citron, supra note 10; Keegstra, supra note 43.
115 For a more complete discussion of zoning software applications, see Geist, supra note 90.
116 See e.g. Digital Envoy, Press Release “Digital Envoy Announces NetAcuity 3.0 Geo-
Intelligence Solution: Flexible Data Modules Offer Increased Targeting Depth” (29 April 2002),
online: Digital Envoy
117 See Ariana Eunjung Cha, “Rise of Internet Borders Prompts Fears for Web’s Future” The
Washington Post (4 January 2002) E01, online: The Washington Post
118 See Jack Goldsmith, “Against Cyberanarchy” (1998) 65 U. Chicago L. Rev. 1199. It has,
however, been rightly pointed out that the effective enforcement of territorially-based laws would
require content providers to be aware of the laws in every country and to tailor their content to meet
them. As Goldsmith indicated, though, businesses operating in multiple jurisdictions are required to,
and do, familiarize themselves with local regulations and refuse to comply with them at their peril.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
primarily motivated by economic return and thus be adverse to negative publicity, as
well as expenditures for defending against prosecution for violation of the laws of
other jurisdictions. 19 Hate speakers may be less motivated by economic returns and
more interested in self-aggrandizement and attracting a following, goals that may well
cause them to disregard the potential legal consequences of their actions, particularly
to the extent that they understand that the First Amendment protects them.
Like filtering, zoning has been squarely criticized for technical reasons.
Technologies designed to circumvent geolocational technology abound, 2 ‘ not to
mention logical flaws identified in certain software.’ At a more philosophical level,
zoning has been criticized for contributing to the balkanization of the Internet, by re-
enabling enforcement of territorially based laws, a result inconsistent with the original
anarchic sentiments of many Internet stakeholders.’ Perhaps of more concern are the
risks to privacy presented by these technologies, to the extent that they facilitate the
often unauthorized collection of pools of potentially quite personal data, including
user location.’23
B. Acceptable Use Policies
A number of ISPs require their subscribers to abide by acceptable use policies
(“AUP”s), some of which include expected standards respecting on-line behaviour
possibly applicable to hate propaganda. These kind of provisions have been primarily
119 The perceived risk of prosecution may be greater in the context of US-based on-line gambling
services, since gambling laws vary from state to state, but may well be enforceable against residents in
other states.
120 See “Putting It in Its Place”, The Economist (11 August 2001) at 18, online: The Economist
computers in other countries, encrypting or anonymously sending their e-mail, and/or using software
to cloak their on-line identities.
121 See Anick Jesdanun, “The Potential and Peril of National Internet Boundaries” The San
Francisco Examiner (4 March 2001), online: The San Francisco Examiner
122 See Joel R. Reidenberg, “The Yahoo! Case and the International Democratization of the
Internet”, Fordham Law & Economics Research Paper No. 11 (April 2001), online: Social Science
Research Network Electronic Library
regulation of Internet activity acts as a democratizing force, returning decisions about what rules
should govern from an elite technocracy to the officials of sovereign nations.
123 Publicly compelled zoning at the national level is implausible, as Canada is a single zone with
respect to current restrictions on hate propaganda. As such, it would make little sense for the Canadian
government to require hate speakers within the country to screen out Canadian users seeking access.
At the international level, while jurisdictions could agree to co-operate in cross-border zoning, it is
unclear whether such measures would survive US First Amendment scrutiny, particularly in the hate
propaganda context. However, at least one member of the US Supreme Court has not ruled out the
possibility that publicly mandated zoning on the Internet may be constitutional (see Lessig & Resnick,
supra note 112).
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
85
motivated by efforts to immunize these businesses and organizations from liability
that may flow from certain types of on-line conduct by their customers (particularly
defamation).1 24 One US court has suggested
that, from a First Amendment
perspective, AUPs may be preferable to government regulation for achieving the
public policy objective of limiting young library patrons’ access to harmful content. 25
However, as the following analysis of AOL’s AUP demonstrates, AUPs are at best
likely to play only a limited role in restricting Internet hate propaganda.
1. The America Online Model
AOL is a provider of on-line services, including e-mail and chat rooms. To
subscribe for its services, customers must agree to AOL’s “Terms of Service”, which
include a “Membership Agreement” and “Community Guidelines” (“Guidelines”)
that outline rules and expected standards of on-line behaviour. The Terms of Service
specify that AOL does not pre-screen content posted to its site and reserves for itself
the decision to remove or block access to harmful or offensive Internet content.
Violation of the Guidelines may result in AOL issuing a warning to the member or
unilaterally terminating the member’s service.126
The Guidelines begin by stating:
Like any city, we take pride in-and are protective of-our community. That’s
why our community standards are important. Communities of all sizes rely on
civic pride and the duty of all citizens to help with things like picking up litter,
getting out of the way of ambulances, reporting crime, and abiding by the law.
These Community Guidelines tell you what you can expect from AOL, as well
as the kind of on-line behavior we expect of you. 127
They specify that AOL takes no responsibility for content that is not removed,
and explicitly state that “the AOL Community Guidelines … including AOL’s
enforcement of those policies, are not intended to confer, and do not confer, any
rights or remedies upon any person.’ 12
‘ As a result, AOL reserves its right to remove
the content, warn members, and/or terminate their service where their behaviour is
inconsistent with community standards, but does not confer on members any legal
right to insist on enforcement against other members. Nevertheless, the Guidelines
124 Their aggressive implementation has been described as an attempt to “manage, avoid and/or
mitigate every potential risk imaginable” (Karen K. Harris, “Issues for Healthcare Companies When
Contracting with ASPs” (2001) 19 J. Marshall J. Computer & Info. L. 569 at 584). Note, however,
that ISPs in
in certain situations. See
Telecommunications Act of 1996,47 U.S.C. 230.
the US are also statutorily protected from liability
125 American Library, supra note 93.
126 AOL, “Member Agreement”, online: AOL Legal Department
127 AOL, “Community Guidelines”, AOL Legal Department online:
128 AOL, “Member Agreement”, supra note 126.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
request that members report content that may violate the standards and state explicitly
that “[h]ate speech is never allowed,”’29 specifying, in material part, that service may
be terminated for members who:
Harass, threaten, embanass, or do anything else to another member that is
unwanted. This means: don’t say bad things about them, don’t keep sending
them unwanted Instant Messages, don’t attack their race, heritage, etc. If you
disagree with someone, respond to the subject, not the person.
Transmit or facilitate distribution of content that is harmful, abusive, racially or
ethnically offensive, vulgar, sexually explicit, or in a reasonable person’s view,
objectionable. Community standards may vary, but there is no place on the
service where hate speech is tolerated.130
The Guidelines articulate a “zero tolerance policy” with respect to illegal
behaviour by members, warning them that AOL will terminate their service and
cooperate with law enforcement authorities. Members are admonished to obey off-
line rules, including the law of foreign jurisdictions, when on-line. Again, however,
this aspect of the Guidelines is not couched in language that would permit members
to require AOL to enforce it.
The Guidelines for members are buttressed by terms and conditions of use
purportedly binding all users’ 3′ of any communication services offered through
AOL’s site, including chat rooms, message boards, and newsgroups, to its rules of
user conduct (“Terms and Conditions”). 32 The Terms and Conditions prohibit using
AOL’s site for an unlawful purpose, including uploading, posting, or otherwise
distributing or facilitating distribution of any content that is, among other things,
unlawful or threatening, or that “victimizes, harasses, degrades, or intimidates an
individual or group of individuals on the basis of religion, gender, sexual orientation,
race, ethnicity, age, or disability.”’33 While the Terms and Conditions reiterate that
AOL does not generally pre-screen, monitor, or edit the content that may be available
through its site, they reserve to AOL the sole discretion to remove any content that, in
its judgment, does not comply with the Terms and Conditions or is “otherwise
harmful, objectionable, or inaccurate.”‘ 34
129 AOL, “Community Guidelines”, supra note 127.
130 ibid.
131 The term “user”, as opposed to “member”, presumably is intended to secure “agreement” as to
AOL’s discretion with respect to content from non-members, who are not bound by the Membership
Agreement and the Guidelines.
13 2 AOL, “AOL.com Terms and Conditions of Use”, online: AOL.com
… AOL, “Agreement to Rules of User Conduct”, online: AOL.com
134 ibid.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
87
AOL has an administrative mechanism through which complaints relating to
content are reviewed and decided upon.’35 While providing a potential avenue for
private enforcement of restrictions on hate propaganda emanating from AOL’s site in
the US, the language of the Guidelines and Terms and Conditions minimizes the
opportunity for consistent enforcement.
2. Limitations on Effective Restriction Through AUPs
Effective restriction of the dissemination of Internet hate propaganda from the US
through private enforcement of relevant provisions in AUPs, such as AOL’s, is
limited in that they are drafted to maximize flexibility for ISPs. Provisions restricting
on-line conduct can be expected to be drafted to impose obligations on users, while
preserving the discretion of the ISP to determine both whether those obligations have
been satisfied and whether any penalty is appropriate in the circumstances.’36 AOL’s
AUP provides an important illustration of this point.
Unlike many other ISPs, AOL is certainly not known for taking an anarchic,
“anything goes” approach to on-line conduct. In fact, it has been criticized for taking
quite the opposite approach to create a very “managed” community. In particular,
critics note AOL’s purported attempt to extend its Terms and Conditions beyond its
subscribers to the wider body of Internet users that may access its on-line services.’
Even AOL’s AUP, however, preserves significant discretion for AOL to determine
what constitutes unacceptable behaviour and whether to impose penalties. AOL has
made clear that it does not consider itself bound by its Terms of Service to take any
action at all in relation to hate speech on its site, and was successful in obtaining a
court ruling to that effect.’
135 For an insider’s perspective on the functioning of this internal mechanism, see Rita Ferrandino,
‘illage Voice (27 March
“Terms of Service: Sweaty Scenes from the Life of an AOL Censor” The
2001), online:
136 Further, the anti-regulation culture prevalent among many ISPs, particularly smaller independent
ones, as well as often limited financial resources, may lead them not to impose AJPs or, at minimum,
to ensure they have plenty of scope to avoid imposing penalties in relation to conduct-based
restrictions.
137 See e.g. Janelle Brown, “Velvet Rope Bolts over AOL ‘Censorship”‘ Wired News (23 July 1997),
online: Wired
on the Intemet: Regulating Web Authorship by Students” (2000) Computer L. Rev. & T. J. 61.
138 See Noah, supra note 88 at 535, in which the court summarily dismissed a former AOL
member’s claim that AOL had violated its Terms of Service and was discriminating in the provision of
a public service by refusing to remove anti-Muslim slurs made in two AOL chatrooms between 1999
and 2001. Although the court dismissed the claim, its reasons confirmed Noah’s position that the
comments complained of were “offensive, obnoxious and indecent”. It is possible, however, that ISPs
could be held liable for false or deceptive trade practices if they fail to deliver on promises of “hate-
free” environments made in their AUPs and other marketing material: Federal Trade Commission Act,
15 U.S.C. 45.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
is not limited
to contractual
C. US-Based Self-Regulatory Organizations
Private regulation by Internet intermediaries
restrictions implemented by individual ISPs. ISPs, frequently in efforts to stave off
public regulation, have formed self-regulatory organizations. Some of these
organizations impose codes of conduct that require members to take various levels of
action in relation to illegal and offensive content emanating from their subscribers,
including co-operating with hotlines receiving reports from Internet users. While
many ISP organizations are territorially based in particular jurisdictions around the
world, some have formed co-operative
from other
jurisdictions. These organizations, again, present the prospect of relying on private
actors to assist in restricting Internet hate propaganda. Given the focus of this paper
on mechanisms for addressing hate propaganda emanating from the US, this section
will focus on US-based organizations.
links with organizations
1. ISP Organizations in the US
New Commerce Communications reported, as of June 2003, three national and
thirteen state ISP organizations in the US.’3 9 Of the national associations, only one-
the United States Internet Service Providers Association (“USISPA”)-provides
direct guidance to members with respect to content regulation, and none provides a
code of conduct for members. Of the thirteen state ISP organizations, it appears that
only two-Florida and Virginia-even tangentially address the issue of content. The
Florida ISP Association provides members with a specific code of ethics that
encourages compliance with laws and the reporting of on-line obscenity through a
hotline. 4 The Virginia ISP Alliance provides standards of excellence for members,
which encourage them to maintain free expression and privacy. 4′ Although the
USISPA is the American association most directly addressing content regulation
139 These national organizations include the Internet Service Providers’ Consortium, the American
ISP Association, and the USISPA. State organizations are reported in California, Colorado, Florida,
Iowa, Missouri, Nevada, New Hampshire, New Mexico, Ohio, Texas, Virginia, Washington, and
Wisconsin. See New Commerce Communications, “The Internet’s Most Complete Site of ISP
Business Organizations and Associations” (2003), online:
ISPs as its members: online: NYSTA
40Florida Internet Service Providers Association, “Code of Ethics” (2003), online:
Web site, which provides a mechanism for reporting potential online violations of US obscenity laws,
and for requesting their investigation by a US Attorney General. The site indicates that US attorneys’
offices received 2,232 reports of on-line obscenity in May 2003 and a total of 24,929 in the period
from June 2002 through May 2003 (Morality in Media, Inc., “Number of Reports to US Attorneys’
Offices June 2002 thru May 2003” online: obscenitycrimes.org
141 Virginia ISP Alliance, “Standards of Excellence” (2003), online: Virginia ISP Alliance
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
89
(albeit very vaguely), neither it, nor any of the other US-based associations require
their ISP members to report on their activities with respect to restricting or reporting
offensive or illegal Internet content.
The USISPA was formed in January 2002 by several large ISPs, including AOL,
Verizon, and Earthlink, to focus on issues of significance to large service providers,
law enforcement
including obligations relating
initiatives. 14 2 Rather than implementing codes of conduct with respect to individual
the USISPA has articulated
members’ addressing illegal or offensive content,
“founding principles” relating to Internet content. Unsurprisingly, these principles
reflect a desire to maximize flexibility for ISPs in relation to content, while
immunizing them from liability for that content. The principles state:
to domestic and
international
As a general rule, liability for Internet content should rest with the
creator or initiator of the illegal content and not with an entity that
retransmits, hosts, stores, republishes, or receives such content.
When serving as conduits for Internet traffic in transit, ISPs should
have no liability or responsibility for the content of such traffic. However,
ISPs should have the right to block or filter such traffic in order to protect
the interests of the ISP or others and should obtain “Good Samaritan”
immunity from liability for such action.
When serving as the hosts for Internet content, whether on a website,
news group, chat room, third-party transaction site or other application,
ISPs should have no liability for content that was not created by the ISP.
But ISPs should accept responsibility for disabling access to such hosted
content in accordance with procedures in an applicable law, such as the
Digital Millennium Copyright Act, or a court order. ISPs should retain the
right to voluntarily disable access to content that they host and should
obtain “Good Samaritan” immunity from liability for such action.
An ISP’s obligation to disable access to hosted content should be
based upon: (a) clear and specific identification of the content in question;
(b) the technical and economic feasibility of disabling access to the content
in question; and (c) a legal framework or court order establishing the ISP’s
obligation to disable access to the content in question, and immunity from
liability if they do so.
*
As a general rule the law applicable to Internet content should be the
law of the jurisdiction in which the content is stored. Such a rule creates
predictability and incentives for safe and secure electronic commerce. 143
Further, although the USISPA is allied formally with the European Internet
Service Providers Association (“EurolSPA”), an umbrella group of ISP associations
(“ISPAs”) from throughout the EU, it would not appear that participation in the
142 David McGuire, “ISP Giants Form New Lobbying Group” Newsbytes (14 January 2002), online:
USISPA
14 USISPA, “USISPA Founding Principles”, online: USISPA
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
alliance requires any particular form of content regulation by USISPA members.”
While there is no indication that members of the USISPA are required to co-operate
with hotlines situated in either the US or the EU, the US is an associate member of
the Intemet Hotline Providers Association of Europe (“INHOPE”), a co-operative of
EU based hotlines focusing largely on on-line child pornography.’45
2. Limitations on Effective Enforcement Through US ISPAs
its broad statements about content
There appears to be no internal mechanism for enforcing USISPA member
compliance with
regulation. Indeed, an
enforcement mechanism would make little sense, given the breadth of discretion
reserved to members with respect to content regulation. Nevertheless, the USISPA’s
decision to ally itself with EuroISPA and its constituent members may create an
opportunity for pressuring USISPA members to comply with reasonable requests to
remove offending material, either through hotlines or direct requests between
associations. Restricting dissemination of Internet hate propaganda through existing
hotline arrangements, however, may be of limited efficacy in the short-term in two
respects. First, hotlines currently focus predominantly on child abuse and child
pornography, rather than hate propaganda, except in cases such as the UK where on-
line racist incidents are reportable. 46 Second, to the extent that US-based hotlines
such as www.obscenitycrimes.org focus on reporting content that may be illegal in
the US, most forms of hate propaganda are unlikely to be addressed.
D. Ad Hoc US ISP Responses to Extra-Territorial Public Policy
US-based ISPs have voluntarily undertaken on an ad hoc basis to address content
considered illegal in other jurisdictions. In the first case, US-based ISPs have
voluntarily agreed to proactively monitor and take measures to disable access to
Internet content likely to be deemed illegal or offensive according to the public policy
of the nation in which they are operating, albeit without the express benefit of a
public determination with respect to the legality of the content. In the second
situation, US-based ISPs have removed Internet content from US servers following
court or tribunal pronouncements of illegality in other jurisdictions. While the latter
type of response better ensures public accountability and transparency, both
demonstrate
that regardless of issues of legal enforceability, public regulation
continues to play an important role in regulating Internet content by guiding private
decision-making.
144 The precise terms of the memorandum between the USISPA and EurolSPA have not been
located. However, even with respect to its EU members, EuroISPA has left content regulation to each
national association. See online: EurolSPA
145 See online: INHOPE
146 See Internet Watch Foundation, “Internet Watch Foundation Annual Review 2002”, online:
Internet Watch Foundation
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
91
1. “Pledging” Proactive Observance of “Local” Laws
the state regulations on Internet
Three hundred ISPs operating in China, including US-based Yahoo!, 147 recently
signed a “Public Pledge on Self-Discipline” (“Pledge”), 48 which requires them to:
“abide by
information service management
conscientiously [and refrain from] producing, posting or disseminating pernicious
information that may jeopardize state security and disrupt social stability … [or]
establishing links to Web sites that contain harmful information.”‘ 149 In fulfilling these
obligations, members are expected to “inspect and monitor the information on …
domestic and foreign websites … [and] refuse … access to those Web sites that
disseminate harmful information.”‘ 5 Prospective violations of the Pledge are
reportable to an executing agency, vested with authority to investigate and publish
findings, as well as to revoke membership in some circumstances. 5’
The Pledge seems to require ISPs to make private decisions as to the legality in
China of content that they host or to which they provide access. The Pledge itself is
obviously not a mechanism for addressing Internet hate propaganda flowing into
Canada. Further, it is limited in scope to the operations of US-based ISPs outside of
the US. It does not explicitly impose a duty to eliminate offensive or illegal content
from servers located within the US, but simply to work toward disabling access to
that content in China. Nevertheless, it presents a model to be considered in terms of
US-based ISP operations within Canada. It is suggested, however, that voluntary
agreements
to public
determinations, described in subsection 2, are preferable with respect to public
accountability.
reactively address hate propaganda
to
in
response
2. Private Enforcement in Response to Public Decisions
A French court ordered US-based ISP, Yahoo!,
to take measures to prevent
French citizens from accessing on-line auctions of Nazi memorabilia. 5 2 The
memorabilia included items featuring swastikas and other Nazi-related symbols that
violate French penal code restrictions on racist content. Even before Yahoo!
successfully challenged the enforceability of the French court order under the First
147 See “Yahoo’s China Concession” Editorial, The Washington Post (19 August 2002) A12, online:
The Washington Post
14′ Digital Freedom Network, “Pledging Self-Discipline” (1 April 2002), online:
149 Digital Freedom Network, “Public Pledge for Self-Discipline for China Internet Industry” (1
April 2002), art. 9, online: Archive.org
” Ibid., art. 10.
” Ibid., arts. 22-23.
152 League Against Racism and Antisemitism v. Yahoo! Inc., County Ct. of Paris, 20 November
2000, No. RG: 00/05308, online: Center for Democracy & Technology
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
Amendment,’53 it had already announced that it would “no longer allow items that are
associated with groups which promote or glorify hatred and violence, to be listed on
any of Yahoo’s commerce properties.” ‘154 The privately imposed ban, while arguably
narrower in terms of the content limited, was clearly broader in the scope of its
application than the restriction ordered by the French court. Yahoo! banned the
related content altogether, rather than limiting the ban to French citizens. Yahoo!’s
public position was that its decision was motivated by pressure from human rights
groups such as the Simon Wiesenthal Center and the Anti-Defamation League, rather
than by the French court decision.’55 However, the timing of Yahoo!’s policy makes it
difficult to imagine that French public regulation had no impact on its decision to
impose a ban.
Subsequently, the ISP Qwest’56 agreed to terminate service to the “Zundelsite”, a
Web site including contents that the CHRT in Citron found violated subsection 13(1)
of the CHRA. The CHRT noted that the practical impact of its cease and desist order
would be limited, given that the respondent no longer resided in Canada and that “the
technology involved in the posting of materials to the Internet … arguably makes it
much easier to avoid the ultimate goal of eliminating the material from telephonic
communication.”‘ 57 While there was no attempt to enforce the cease and desist order
against the related content provider or his ISP in the US, the CHRT’s public
denunciation of the Web site encouraged Qwest to terminate service to it. A Qwest
official stated that it removed the site after receiving notice of the CHRT’s decision.
Qwest’s AUP prohibits distribution of hateful, obscene, abusive, or excessively
violent material.’58 As with the Yahoo! case, it would appear that the denunciatory
impact of a public determination of illegality influenced the exercise of private
discretion in accordance with public policy outside of the US.’59
The Yahoo! and Qwest examples demonstrate
that public regulation and
decisions made pursuant to it can serve as effective guides for private enforcement
mechanisms. Private decision-makers are relieved of the obligation to assess the
legality of any particular content and are able to point to the law, rather than a private
taste for censorship, as the basis of their decisions. 6 While the law may influence
153 Yahoo! v. Ligue, supra note 71.
114 Troy Wolverton & Jeff Pelline, “Yahoo to Charge Auction Fees, Ban Hate Materials” CNET
News.con (2 January 2001), online: CNET
155 See Lisa Guernsey, “Yahoo to Try Harder to Rid Postings of Hateful Material” The New York
imes (3 January 2001) C2, online: The New York Times
156 See online: Qwest
157 Citron, supra note 10 at para. 298.
158 See Adrian Humphreys, “U.S. Internet Giant Pulls Zundel’s Web Site: Canadian Rights Panel
Warned Firm of Hate Literature” National Post (13 May 2003) A9.
159 This example demonstrates the practical limitations of ad hoc private responses following public
decisions. The Zundelsite was back up and running using another host server by 12 May 2003 (ibid.).
160 These examples illustrate Sunstein’s argument that some market players who do not wish to
discriminate rely on regulation to eliminate the choice about whether to cater to demands for
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
93
and shape market behaviour and norms, the influence of demands by vocal non-
governmental human rights groups on the exercise of private discretion in these
situations should not be underestimated. 6′ The ad hoc and purely voluntary nature of
these kinds of private enforcement choices make them unreliable mechanisms for
protecting important matters of public policy. Moreover, in the absence of public
reporting, it is difficult to gauge either the breadth of these private practices or their
consistency with public policy.
for restricting
E. Private Regulation Is Not a Substitute for Public Regulation
Private regulatory mechanisms present an opportunity
the
dissemination of hate propaganda from the US., given the general absence of First
Amendment scrutiny of private action. In addition, purely private regulation may be
more flexible than state regulation and thus better able to react quickly to new
phenomena, such as transmission through Internet applications other than the
relatively accessible WWW.’62 However, practical factors associated with each of
these mechanisms
in restricting the flow of Internet hate
propaganda emanating from the US. More pressingly, there are broader policy
concerns as to the wisdom of relying too heavily on the private market to regulate
harmful discriminatory behaviour. The question then becomes whether the greater
breadth and peer pressure that may be associated with group-based self-regulation
through codes of conduct can be harnessed to encourage individual ISPs to privately
regulate in accordance with the policy goals articulated in national public regulation
and international human rights instruments.
limit their efficacy
1. Practical Limits
The current state of the art limits the ability of filtering and zoning technologies
to effectively restrict Internet hate propaganda. The labelling of content necessary to
discriminatory content in the marketplace: Cass Sunstein, Free Markets and Social Justice (New
York: Oxford University Press, 1997) at 154 [Sunstein, Free Markets].
161 The Simon Wiesenthal Center, the ADL, and BiasHELP (an anti-discrimination group) appear to
have had significant impact on the policies of individual ISPs and on-line content providers in
connection with regulating racist expression on-line. See e.g. Troy Wolverton, “eBay Asked to Pull
KKK Items from Site” CNET News.com (2 February 2000), online: CNET
Auctions” CNET News.com (3 February 2000), online: CNET
2002), online: Japan Today
162 See Sabine Frank, “Co-operative Forms of Regulating the Intemet” (Report at the Council of
Europe Forum on Harmful and Illegal Cyber Content: Self-Regulation, User Protection and Media
Competence, Strasbourg, 28 November 2001), online: Council of Europe
(2002) 3 Colum. Sci. & Tech. L. Rev. 1.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
enable filtering, even if restricted to WWW content, presents a daunting task. First
Amendment protection from the risk of prosecution and penalties means that US-
based hate speakers are unlikely to be motivated to accurately label or to block access
to their message, regardless of whether that message is illegal in the jurisdiction
where it is accessed. The prospective efficacy of filtering and zoning is further limited
by relatively effective technological circumvention measures. Finally, filtering and
zoning at the ISP level raise concerns with respect to monitoring and recording of a
potentially highly private body of user data by ISPs.’63
Private regulation through AUP’s or ISP self-regulation, as currently structured in
the US, is also limited by practical difficulties. AUPs, even those drafted by
interventionist ISPs such as AOL, reserve maximum flexibility for ISP discretion and
judgment. Further, while there are instances of individual ISP regulation of content in
response to public determinations of illegality outside of the US, in the absence of a
more organized and concerted effort to pressure for compliance, this presents an ad
hoc avenue at best. Although US-based ISP organizations could act as focal points for
asserting international pressure to self-regulate Internet hate propaganda, the absence
of enforceable codes of conduct among US-based organizations and the current focus
of hotline efforts on on-line child pornography limit the likelihood of restriction
through this form of private regulation, at least in the short term.
Identification of these practical limitations, however, is not intended to suggest
that private regulation through the implementation of technological measures, AUPs,
or self-regulatory organizations should necessarily be rejected. Rather, it is intended
to highlight that, despite the general absence of a First Amendment barrier, private
regulation, like public regulation, offers at best a partial solution for restricting the
extraterritorial flow of Internet hate propaganda emanating from the US. Of more
fundamental concern is whether primary or substantial reliance on private market
solutions in the context of hate propaganda would divest public authorities of their
responsibility to safeguard Canadian public policy, with no degree of certainty that
private market
international collective
commitments to equality and diversity.
responses would serve national and
2. Policy Issues
Privatization or shifting responsibilities and power previously within the purview
of the state to private actors is an expanding trend in Canada, the US, and indeed,
163 It is not suggested that privacy related issues completely undermine the concept of ISP-based
filtering or zoning by individual content providers. However, such issues do merit attention to ensure
that users are aware of and provide some form of meaningful consent to any filtering or zoning
technologies privately implemented. Current privacy protection laws should be reviewed to determine
their adequacy in addressing issues raised by the implementation of these technologies. See Geist,
supra note 90.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
95
internationally.”6 As jurisdictions become increasingly interconnected economically
and technologically, academic commentators have noted the push toward competitive
advantage through the purported efficiencies of an unregulated private market, even
in areas such as family and human rights law.’65 Many Internet stakeholders and
theorists have urged governments in connected jurisdictions around the world to take
a similar approach to the regulation of the Internet and its content, with some measure
of success.’66 In fact, some ISPs and their self-regulatory organizations openly admit
that moves toward self-regulation were designed to stave off public regulation and
capitalize on what they see as the efficiency and flexibility of private enterprise.’67
Perhaps the most disturbing aspect of this trend is the shift in responsibility for
collective goals away from the state to the invisible hand of the private market, often
generating calls for those historically most disempowered by a free market system to
take action to protect their own interests.
Arguments in favour of individually implemented filtering and zoning reflect this
shift, admonishing individuals offended by particular content to take private measures
to avoid it. However, as suggested above, this approach completely misses one of the
fundamental commitments underlying Canadian restrictions and international provisions
relating to hate propaganda: the broader risk to equality and multiculturalism associated
with the adoption of discriminatory attitudes by those not offended by hateful messages.
Further, the private market has an unimpressive historic record in correcting
discrimination based on personal characteristics such as race, gender, and sexual
identity. 68 Unfortunately, this record is arguably consistent with rational behaviour by
suppliers in seeking to meet consumer preferences. To the extent that there is demand
for hateful content sufficiently widespread to sustain profitable economic activity,
rational suppliers acting in their own economic self-interest will supply product to
’64 The presence of multinational corporations in physical jurisdictions around the globe, and their
seeming immunity to both domestic and international law, has already been the subject of substantial
commentary and concern. For an overview, see Trevor Farrow, “Globalization, International Human
Rights Law and Civil Procedure” (2003) 41 Alta. L. Rev. [forthcoming].
165 See Brenda Cossman & Judy Fudge, eds., Privatization, Law, and the Challenge to Feminism
(Toronto: University of Toronto Press, 2002); Sarah Krieger, “The Dangers of Mediation in Domestic
Violence Cases” (2002) 8 Cardozo Women’s L.J. 235.
166 See e.g. U.S., Department of Commerce, Management ofInternet Names and Addresses (S. Doc.
No. 980212036-8146-02), online: National Telecommunications and Information Administration
Telecommunications Commission, Broadcasting Public Notice CRTC 1999-84, “New Media” (17
May 1999), online: CRTC
In telling
contrast, large media companies have not hesitated to insist on government intervention to protect
copyright interests. See Lawrence Lessig, The Future of Ideas (New York: Random House, 2001).
167 Frank, supra note 162.
168 The historic failure of the private market with respect to racist propaganda, and the devastating
results that flowed from this failure are chronicled in Alexander Tsesis, “Prohibiting Incitement on the
Internet”(2002) 7 Va. J.L. & Tech. 5.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
meet that demand.’69 There is no reason to expect that economic actors in the Internet
market will behave any differently. To the extent that there exists a sufficiently
widespread demand for discriminatory content among Internet consumers, or indeed,
that the Internet can be used to stimulate demand, economically rational suppliers can
be expected to meet it.
Additionally, in the context of the Internet, the economic rationality behind
supplying discriminatory content combines with relatively widespread idealization of
the Internet as a wide open, anarchic marketplace of ideas, to work against effective
restriction of Internet hate propaganda solely through private regulation. ISPs
considering private regulation of hate propaganda will have to weigh the potential
reputational effects of continuing to carry hateful content versus being branded as
censors interfering in the operation of the marketplace. Unfortunately, economic
rationality may dictate that some ISPs will refuse to restrict content in relation to the
least powerful members of society, reasoning that these groups are least able to bring
economic pressure to bear. 7 ‘
Assuming an economic incentive to supply hateful content, what then has
prompted some ISPs to impose private restrictions on Internet hate propaganda? In
the US, it is almost certainly not the threat of prosecution, imposition of legal
penalties, or impending public regulation, given First Amendment protections. These
responses may be attributable to concerns that demand for certain types of hate
propaganda is not sufficiently widespread to merit the risk of reputational harm that
may arise from being associated with its continued supply. However, in the absence
of some mechanism for public scrutiny, decisions on whether to restrict this type of
content are unlikely to be affected by reputational concerns. It is in this regard that
continued public regulation and more widespread self-regulatory organizations with
the ability to assert peer pressure can perhaps be of greatest efficacy by publicly
exposing what are currently private decisions. 7’
169 For a more complete articulation of the inability of the market to address discrimination based on
personal characteristics, such as race, that are otherwise irrelevant to decisions such as hiring and
promotion, see Sunstein, Free Markets, supra note 160 at 151-54.
170 For example, in the aftermath of the 11 September 2001 attacks, persons of the Islamic faith have
been increasingly targeted by on-line messages of hate. These messages may, however, carry very
little weight in terms of the rational economic decision-making of ISPs, given what appears to be a
relatively broad-based vitriol toward persons of Middle Eastern and Asian dissent in the US, and
given the racial profiling and targeting of members of these groups resulting from legislative and law
enforcement initiatives aimed at terrorism. See Salah D. Hassan, “Arabs, Race and the Post-
September 11 National Security State” Middle East Report 224 (Fall 2002), online: Middle East
Report
Providers Curb Both Terrorist Postings and an Anti-Islamic Backlash” The New York 7imes (17
September 2001) C8, online: The New York Times
17 Likewise, public accountability will also be important to ensure that private actors do not restrict
expression beyond what is democratically necessary, for example, by restricting criticism of
government action. See generally Lessig, supra note 1.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
97
Part VII briefly comments on the question “where do we go from here?” first
exploring the goals we might wish to achieve in relation to existing private regulation
and then turning to the more practical issues of who might assist in achieving them
and how they might be advanced.
VII. Organizing Private Action to Work Toward Public Goals
A. Goals for “Publicizing”172 Private Action
The suggestions in subsection B below as to who might assist with organizing
private action in service of national and international human rights objectives, and
how they might go about doing that, are motivated by two goals. First, although
public regulation of human rights issues such as hate propaganda is generally
preferable, given First Amendment constraints on state enforcement of public
regulation, we ought to explore ways to supplement public regulation by capitalizing
on and expanding existing examples of private regulation. Second, in doing so, we
ought to look for ways to make private regulation more systematic, transparent, and
publicly accountable.
With these goals in mind, subsection B turns to the question of how the project of
publicizing private action in a more systematic fashion might be advanced, having
regard for the wisdom gained with respect to UN and Organization for Economic
Cooperation and Development (“OECD”) efforts to address multinational corporate
compliance with, among other things, internationally proclaimed human rights.’73
B. Advancing the Project
1. Who Might Be of Assistance?
There are at least four groups that might be individually or collectively involved
in advancing the project of mobilizing private regulation in service of internationally
172 For a discussion of how privatization might actually be used to extend, rather than retract, public
norms and policy, see Jody Freeman, “Public Values in an Era of Privatization: Extending Public Law
Norms Through Privatization” (2003) 116 Harv. L. Rev. 1285.
173 Although
there are a myriad of international
initiatives to encourage private corporate
compliance with international standards, norms, and law, this paper relies upon and refers primarily to
two: The Global Compact and the OECD Guidelines for Multinational Enterprises. See United
Nations, “Overview: What
the Global Compact?”, online: UN
on International Investment and Multinational Enterprises, The OECD Guidelines for Multinational
Enterprises: Text, Commentary and Clarifications, Doc. No. DAFFE/IME/WPG(2000)l5/Fina 2001),
online: OECD
is
MCGILL LAW JOURNAL / REVUE DE DROITDE MCGILL
[Vol. 49
defined human rights 7 4 reflected in the Canadian Provisions: collective international
organizations such as the UN and the OECD, national governments, ISPAs, and
independent human rights NGOs.
International collectives such as the UN and the OECD are ideally suited to assert
a considerable degree of clout that is likely to be of assistance in encouraging private
actors to comply with international norms and standards. While both organizations
are concerned with addressing human rights on a global scale, the UN may be better
situated to address hate propaganda,’75 given its related committees’ and rapporteurs’
oversight powers in relation to compliance with the CERD, the ICCPR, and the
Universal Declaration of Human Rights.’76 While the UN is thus arguably well
placed to address this human rights issue, the OECD’s developed expertise on issues
of Internet regulation should be capitalized upon.’77
to encourage US ISPs
Involving national governments
to comply with
internationally established human rights that animate legislation like the Canadian
Provisions offers the important element of additional clout to the protection of these
rights and the potential for
legal mechanisms of enforcement. Government
involvement would forestall criticisms like those leveled at the Global Compact.
NGOs, comparing the Global Compact to the OECD Guidelines, have rightly noted
that national government involvement in monitoring and reporting on compliance
174 Notably, however, any international approach premised on the CERD and the ICCPR will focus
on racist propaganda only, whereas the Canadian CHRA provisions address a broader range of
discriminatory grounds.
175 For example, the UN Commission on Human Rights (“UNCHR”) recognized at its 2001 World
Conference the role of emerging technology both in alleviating and perpetuating racism, racial
discrimination, and xenophobia. While it urged states and the private sector to develop self-regulatory
measures to combat racism, the UNCHR did not assume an active role in implementing or monitoring
such measures. See UNCHR, “Report of the World Conference against Racism, Racial
Discrimination, Xenophobia and Related Intolerance” (Durban: 2001), A/CONF. 189/12, online: UN
176 G.A. Res. 217(111) UN GAOR, 3d Sess., Supp. No. 13, UN Doc. A/810 (1948) 71 provides both
for protection of equality (art. 2) and freedom of expression (art. 19), subject to such “limitations as
are determined by law solely for the purpose of securing due recognition and respect for the rights and
freedoms of others and of meeting the just requirements of morality, public order and the general
welfare in a democratic society” (art. 29).
177 ‘Me OECD has already been involved in facilitating discussion of ISP self-regulation. See e.g.
OECD, BL4C/OECD Forum: Internet Content Self-regulation (Paris: OECD, 1998), online: OECD
guidelines relating to consumer protection in e-commerce. For a 2003 report on these guidelines, see
OECD, Directorate
for Science, Technology and Industry Committee on Consumer Policy,
Consumers in the Online Marketplace: The OECD Guidelines Three Years Later, Doc. No.
DSTI/CP(2002)4/FINAL (2003), online: OECD
Subsequently, the OECD proposed specific Internet content guidelines relating to “spamming”. See
OECD, OECD Guidelines for Protecting Consumers from Fraudulent and Deceptive Commercial
Practices Across Borders (2003), online:
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
99
moves the OECD Guidelines a step closer to binding private enterprise than does the
Global Compact.’78
In the Internet hate propaganda context, it may well be possible to involve
national governments
in monitoring and reporting, as well as enforcement, as
reflected in co-operative efforts such as the Additional Protocol. However, given the
current First Amendment approach, US government involvement is unlikely, thereby
eliminating the participation of a key player. Nevertheless, like-minded countries
including Canada ought to involve themselves in these kind of efforts. In this regard,
the OECD Guidelines present a useful model, requiring adhering states to identify
national contact points obligated to, among other things, monitor and report on
compliance by private enterprise within their jurisdiction and to attend annual
meetings overseen and facilitated by an OECD committee. 79
Unlike the situation in the US, a number of ISPAs in other jurisdictions have
implemented codes of conduct that require their members to address illegal and
offensive content. Some also require co-operation with privately organized hotlines
and hotline coalitions that receive and act on reports relating to offensive and illegal
content, 8 These organizations, and the affiliations among
the
opportunity for more broadly based and consistent private regulation. Like individual
ISP regulatory decisions, however, the implementation of their codes is frequently not
them, present
178 See e.g. National Policy Association, “The UN Global Compact and the OECD Guidelines”,
to Promote Corporate Social Responsibility
P9 OECD Guidelines, supra note 173 at 46-50.
180 The Canadian Association of Internet Providers, for example, sets out a fair practices policy
statement that provides that its members will not knowingly host illegal content or conduct and will
take action when notified of either: Canadian Association Internet Providers, Fair Practices
Document (31 August 2000) at 11, online:
Internet Service Providers Association (“UKISPA”) requires members to take steps to ensure services
and promotional material (although not third party content) do not contain material inciting violence,
sadism, cruelty, or racial hatred and to comply with notification of illegal content issued by the
Internet Watch Foundation (“IWF”). The UKISPA reserves the right to impose a variety of penalties
on members who fail to abide by its code, including membership suspension and termination, as well
as publication of the member’s identity and the ISPA’s findings with respect to any complaint relating
to that member: UKISPA, “Code of Practice” (as amended 19 April 2002), art. 8, online: UKISPA
for the Voluntary Self-Monitoring of Multimedia Service Providers) (“FSM”) in Germany goes
further, requiring members to self-monitor for a variety of categories of impermissible and illegal
content, including content inciting hatred and violence against minority groups and instigation of
racial hatred. Members agree not to provide this content directly or to provide a switch for its use. The
code includes a complaints and hearing mechanism, which can result in the publication of reprimands.
In addition, the code provides a mechanism for reviewing and issuing public reprimands in relation to
non-member ISPs that provide impermissible content or switches for its use in order to encourage non-
members to remove or disable access to that content: Freiwillige Selbskontrolle Multimedia-
Diensteanbeiter,
July 1997), online: FSM
“Code of Conduct”
(9
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
to be uneven at best.’8′
subject to public review and, in any event, appears
Nevertheless, the breadth of these organizations and their potential to harness peer
pressure among members ought to be taken into account when fashioning an
approach to more effectively restrict Internet hate propaganda.
NGOs play an ongoing and important role in monitoring and reporting on failures
by government and private industry to observe basic human rights.’ 82 However, as
one group pointed out in relation to multinational corporate compliance with
international human rights obligations:
it
to suggest
to
is not enough
the assertion of some,
Contrary
that
nongovernmental organizations (NGOs) will assume
this monitoring and
enforcement function. Human Rights Watch has devoted substantial resources
to promoting corporate respect for human rights, but our efforts are just a drop
in the bucket. Neither we nor other NGOs begin to have sufficient resources to
assume an enforcement role that should be the province of governments and the
UN.183
Given the experience of the UN and its related committees and rapportours in
addressing human rights issues, Canada ought to encourage it to become more
directly involved with regulating Internet hate propaganda. While the Global
Compact may have some impact on encouraging private actors in the Internet
industry to comply with international human rights initiatives, by June 2003, few, if
any, had actually “signed on” to participate.’ As such, a more directed effort toward
Internet hate propaganda may be merited, one targeting inclusion of interested
national governments,’85 ISPAs, and existing human rights-based NGOs.
181 Canadian Secretariat, World Conference Against Racism Advisory Committee, Combatting Hate
on the Internet (Issue Position Paper, Hate and New Media Working Group) (31 January 2001),
online: Canadian Heritage
182 See e.g. online: Human Rights Watch
intemet.asp>; Southern Poverty Law Group,
note 161; online: ADL
183 Kenneth Roth, “Corporate Social Responsibility,” Letter to UN Secretary General Kofi Annan,
(28 July 2000), online: HRW
184 The UN reported that by June 2003, 1,018 companies from 56 countries had notified the
Compact office of their intention to participate in the Compact, although it could not advise as to the
status of these companies’ compliance with the Compact’s stated principles. Of the companies listed
as participants, only 45 were from the US, 7 from Canada, and 22 from the UK, compared to 176
from Poland, 86 from India, and 147 from France. From a review of their names, very few appeared to
be ISPs. See United Nations, “Global Compact Participants by Country” (10 June 2003), online: UN
185 Other UN members, such as France, may be highly motivated to participate, even though
participation by the US (and possibly other member states) is unlikely.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
101
2. What Steps Should Be Taken?
As suggested by the Canadian Secretariat to the World Conference Against
Racism Advisory Committee, the UN could establish an independent coordinating
agency to: (i) develop a code of conduct for the Intemet; (ii) monitor and report
annually on Internet hate propaganda; and (iii) develop educational programs to help
governments at all levels to recognize the impact of Internet hate propaganda and the
need for its regulation.’86 The development of a code would need to be done in
consultation with key players, including those listed above.
The monitoring and reporting function and the involvement of ISPs and ISPAs in
the process from the outset would ideally generate an incentive for ISPs and ISPAs,
particularly in the US, to adopt the model. The model should also provide for a
central agency to which individual ISPs or ISPAs report activities relating to the
restriction of offensive or illegal content to foster public accountability and a degree
of transparency. It might also include a notice and takedown system, with notice
being provided to ISPs by the coordinating agency itself or through a designated
hotline responsible for reviewing content in accordance with the strictures of the
ICCPR and CERD, thereby unburdening ISPs of the task of independently assessing
the legality of particular content.
C. Shortcomings
The skeletal framework suggested here is at once an ideal and a compromise. It is
idealistic in the sense that it would require mobilization of international resources at a
time when US resistance to international solutions on the issue has been made
perfectly clear and international tension on issues of human rights and terrorism is
high. Further, it would require expenditure of resources not just by national and
international public agencies, but also by private ISPs and ISPAs, potentially
undermining their willingness to participate.
It is a compromise in the sense that it accepts private regulation and seeks to
build into it a degree of accountability, rather than a direct mechanism for legal
enforcement. Undoubtedly, its non-binding nature and dependence upon what is
effectively moral suasion to encourage private participation and compliance is less
than ideal. As a result, it is likely to attract the participation only of larger private
actors who are concerned about the reputational effects of the monitoring and
reporting suggested, leaving behind a not insignificant group that is impervious to
these effects due to commitments to a vision of an anarchic Internet marketplace or
their non-commercial objectives. Nevertheless, this approach seeks to build on
186 Canadian Secretariat, supra note 181.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 49
experience gained in ongoing international efforts to address transnational problems
and offers the potential for publicizing ongoing private regulation.187
Even if the international mobilization necessary to develop, monitor, and report
on a model code of conduct never comes to fruition, Canada should continue its
efforts to comply with its international obligations by vigorously enforcing the
Canadian Provisions. Public decisions pursuant to that legislation continue to provide
valuable guidance triggering private enforcement against the growing body of
Internet hate propaganda emanating from the US. Further, Canada should sign the
Additional Protocol, joining forces with other nations around the world seeking to
address dissemination of these destructive messages.
Conclusion
as Thierer predicts, allow the US to “[offer]
The export of a First Amendment in code more powerful than that in law may
well,’
the protection of the First
Amendment over the Net to millions of people who have been denied the right to
speak freely in their own countries.” ‘189 However, in Canada, First Amendment
imperialism, particularly on the issue of hate propaganda, holds no freedom-
enhancing promise. Rather, the export of the current US approach to hate propaganda
threatens essential public policy reflected in Canada’s democratically enacted and
constitutionally sanctioned restrictions on this “de-liberating” exercise of private
power. The thinner conception of liberty as freedom from government restriction
underlying the US approach fails to take sufficient account of the de-liberating impact
of hate propaganda on target group members and broader public concerns for equality
and multiculturalism, which are entrenched Canadian constitutional commitments.
Ongoing regulation by private Internet actors, in particular ISPs, could limit the
impact of the current First Amendment approach on Canada’s public policy
objectives. However, private regulation alone is not enough. Achieving the goals of
equality and multiculturalism, and the mutual respect for diversity essential to them,
should not be primarily entrusted
to largely unscrutinized choices made by
individuals in private settings or be foist on target group members by expecting them
to adopt technological means of avoidance. Nevertheless, the comparative advantage
187 Voluntary approaches, initiated in the context of existing public national and international
regulation, have enjoyed some history of success, if on no other basis than raising awareness and
encouraging reporting. See e.g. International Labour Organization, “MNE Declaration” (9 April
2003), online: ILO
reporting on
the
“Tripartite Declaration of Principles Concerning
Multinational Enterprises and Social Policy” (1977), online: 1LO
Johnson, “Public-Private-Public Convergence: How the Private Actor Can Shape Public International’
Labor Standards” (1998) 24 Brook. J. Int’l L. 291.
International Labour Organization’s
188 Lessig, supra note 1.
189 Thierer supra note 2.
2004]
J. BAILEY- RESTRICTION OF INTERNET HATE PROPAGANDA
103
of private regulation with respect to the First Amendment roadblock makes it worth
considering as a regulatory supplement.
international
responses
in numerous
Existing private regulatory practices and structures might be enlisted in service of
public goals, an approach modeled
to
multinational corporations. Key to this effort will be identifying a sufficiently
powerful international oversight body capable of garnering the respect necessary to
encourage private Internet actors to participate and report. If nothing else, monitoring
and reporting requirements would facilitate a more comprehensive empirical
understanding of existing private regulatory efforts, offering the possibility of
increased public accountability and transparency. In this regard, with or without an
international initiative, a Canadian agency such as the CHRC might undertake its
own empirical analysis of ongoing private initiatives.
In the interim, the importance of ongoing enforcement of public regulation
should not be underestimated. Violations of the Canadian Provisions should be
vigilantly prosecuted. Resulting
should publicly denounce hate
propaganda’s inconsistency with Canada’s defining constitutional commitments to
equality and multiculturalism and guide private decision-making regarding restriction
of content currently granted safe haven by the First Amendment.
judgments
