About ECRI


Other CM Decisions

Internal Rules of Procedure

ECRI members

ECRI observers

ECRI in short



Country Monitoring Work

Work on General Themes



Annual reports 



Search (New HUDOC-ECRI database)

Press Releases









Restricted access

Access members

Password reset (expires every two months)


Legal instruments to combat racism on the internet


Terms of reference
Scope of the study
Working procedures

Terms of reference 

On 19 August 1999 the Council of Europe formally commissioned the Swiss Institute of Comparative Law, Lausanne, to produce a report on the legal measures, in particular criminal-law measures, intended to combat racism on the Internet. The study was to be based on the situation in a dozen member countries of the Council of Europe: Germany, Austria, Belgium, Estonia, France, Italy, Norway, Poland, the Czech Republic, Russia, Sweden and Switzerland.

The terms of reference were defined at a meeting between the Deputy Director of the Institute and the European Commission against Racism and Intolerance (ECRI) held in Strasbourg on 15 June 1999. A member of the Institute who attended a meeting of the Internet Sub-Committee of the ECRI held in Paris on 5 November presented a progress report and the object of the research was further refined.

On 31 March 2000, within the agreed period, the Institute submitted the present Report to the ECRI in a bilingual version, partly in French and partly in English.

Scope of the study 

Our first observations in this regard concern the fact that the terms of reference were confined to criminal law: this limitation was justified by the fact that this branch of the law is the best suited to combat hateful words. Having said that, we considered it appropriate to make the occasional reference to civil law or administrative law, which sometimes offer effective means, particularly from the aspect of speed, of ensuring that access to racist sites is blocked, or indeed that these sites are simply closed down.

It will be noted that the study concentrates on legal measures to combat racism on the Internet. The word “legal” must be understood in a broad sense, however, and is not restricted solely to positive law, consisting of legal rules and judicial decisions. In a sphere as mobile and unstable as Internet law (see the caveat below) the majority of countries covered by the study have combined the classic normative approach with the measures deriving from soft law. Although there was no express provision to that effect in the terms of reference, the Institute considered that it could not disregard these more flexible instruments consisting of codes of conduct, ethical requirements, recommendations or hot lines, if not because of their effectiveness, at least because of their strategic importance.

As regards the classic approach, it should be emphasised at the outset that the rules specifically aimed at racism on the Internet – or even more generally at abuse of freedom of expression on the network of networks – are virtually non-existent. Admittedly, there is no shortage of proposals, whether from legal commentators or from the authorities; but in order to avoid overdiversification and confusion (owing to the frequently divergent and contradictory nature of these proposals), we have made a point of dealing only with proposals which are in the process of being adopted; in other words, those being debated in Parliament.

It should further be noted that judicial decisions on the matter are also still rare; and they do not constitute a body of settled case-law, since with few exceptions the only judgments delivered thus far have been those of the lower courts. In order to compensate for this lacuna, the study takes into account judicial decisions delivered in other cases involving unlawful expression on the Internet, in particular pornography and infringement of copyright. The fact that the relevant issues, beginning with the delicate question of the “subsidiary” liability of technical operators (access providers and hosts) means that such an extension of the scope of the investigation is valid, and indeed necessary.

Does this extension of the material scope of the study make up for the reduction of its geographical scope? The reader will note that this report says virtually nothing about the situation in the four central and eastern European countries referred to in the mandate (Estonia, Poland, the Czech Republic and Russia). This relative silence is not due to forgetfulness on our part or to the absence of racist messages on the networks of those countries; in fact, we redoubled our efforts to contact the bodies concerned with racism on the Internet in these countries (prosecution authorities, access providers and human rights organisations) – to no avail. No specific provision in that regard was reported, whether of the classic normative type or soft law1. There is an explanation: the development of the network of networks is still at a very embryonic stage in these countries, where the Internet is still the prerogative of an academic elite.

Finally, the study will also deal with the situation at European Union level. Not that the EU is in the process of mounting a direct attack on the problem of racism on the Internet: that is by no means the case. None the less, the directive on electronic commerce, which is in the process of being adopted, establishes certain standards concerning the liability of technical intermediaries: these standards will be binding on Member States and will have some effect on the fight against the distribution of illegal contents, including racist contents.

Working procedures 

We decided not to present a compilation of reports on a country-by-country basis, since the reader would eventually have been lost in the jungle of national specificities, primarily institutional and procedural. We therefore opted for a horizontal approach, which provides a better means of comparing problems and attempted solutions.

We therefore begin by setting out the various technical and legal difficulties associated with seeking the persons committing offences involving racist expression. We then examine the possibilities of imposing liability on actors other than the person actually committing the offence, first by means of classic legal measures and then by soft law measures. After briefly considering the provisions of international law measures which may be relevant, the study ends by summarising the problem and, in particular, briefly lists the instruments which we consider appropriate or inappropriate. In all cases our general considerations are illustrated by examples of significant developments in one or other of the countries studied.

Before we get to the core of the subject, however, we feel it necessary to describe certain technical data; in particular, it is important to define the role played the various actors involved in the process of disseminating communications on the Internet; similarly, it is worth mentioning the various services offered by the Internet. These differences give rise to nuances and distinctions in the legal regime applicable.

Finally, we must point out that the present study does not review the various criminal provisions of the countries concerned aimed at combating racism in general. That information is available in an earlier report by the Institution, also commissioned by the Council of Europe, entitled Legal measures to combat racism and intolerance in the member States of the Council of Europe, which was published by the ECRI in 1998 and is also available on the Internet2.


The myth of an Internet without faith or law should be dismissed at the outset. This myth of a legal vacuum, which is supported by certain alarmist politicians, amplified by the press and exacerbated by unconsidered declarations of independence by “surfers” eager for absolute freedom, does not stand up to examination. Like any other means of communication, the Internet does not escape the law. As a general rule, the laws governing the right of communication are drafted in a technically neutral manner, which takes into account any dissemination of information irrespective of the medium; consequently, they are fully applicable to messages distributed on the Internet. As we shall see, the problem therefore lies not so much in the absence of adequate material rules as in obstacles to their application in the form of the characteristics peculiar to the network of networks, namely its polycentric structure, its ubiquity and the cover of anonymity.

This is all the more so because, returning to racist expression, our previous report showed that all European countries have at their disposal a more or less effective legislative arsenal to repress hateful expressions. This minimum standard, moreover, is imposed by the United Nations Convention on the Elimination of All Forms of Racial Hatred, Article 4 of which requires the adoption, inter alia, of a provision penalising the propagation of racial hatred outside a strictly private circle. These criminal provisions, which are drafted in general terms, are applicable, inter alia, to hateful expressions disseminated via the Internet.

There is an exception to the common standard, however; it concerns negationism, which means calling in question the existence of genocide. Other than France, where it is an offence to “dispute crimes against humanity”,3 only Switzerland, Germany, Belgium and Austria punish this offence; and in the latter three countries the offence is limited to the denial of genocide committed by the Nazis. This difference in approach is worth mentioning, out, since revisionist sites are flourishing on the Internet.

Finally, the reader’s attention should be drawn to a last point of importance: the changing, or rather ephemeral, nature of the present study. Technology develops very rapidly – who, at the beginning of the 1990s, could have predicted the lightning development of the Internet? – and the law on communications, despite struggling to keep pace, is also developing very quickly. Perhaps the problems described here will no longer be problems in a few years, indeed in a few months; in addition, the solutions recommended are equally volatile. From this aspect, it should finally be noted that the various websites referred to in the footnotes are up to date on 20 March 2000.

1. However, the principal access providers, which are subsidiaries of foreign companies such as IBM, AT&T in Croatia or FREEnet in Russia, refer to the codes of conduct of their parent companies, which in the majority of cases are American.
2. http://www.ecri.coe.int/fr/03/03/f03030001.htm
3. Article 24 bis of the Law of 29 July 1881, as amended by the Law of 13 July 1990.