    |
|
|
|
Freedom of Expression and the Communication NetworksStrasbourg, June 1998 CC-Cult (98) 18 Council of Europe Document prepared for the Electronic Publishing, Books and Archives Project Report prepared by: Department of Information and Library Studies ContentsPart One:
Part Two:
Collections of Internet Policies The author wishes to thank the following for their assistance in various ways, but in each case speaking in their own capacity, not on behalf of their organizations.
Particular thanks to Goff Sargent who provided valuable guidance and support. Part One: The appearance of electronic networks has transformed the environment for the delivery of information to users in ways that deeply concern the players in the traditional Book Economy. There is a commonly expressed sentiment that information and ideas are more than ever before a free currency which cannot and should not be interfered with. At the same time the widespread ability to gain access to, and use, the technology is a source of anxiety to governments, security agencies, defenders of orthodoxies and fixed concepts of public decency which see networks as a threat to their vision of society. The monitoring of electronic communication and the interference with freedom of expression is much discussed and restrictions are frequently proposed. The Internet itself includes many Web sites and newsgroups devoted to this debate. Professional and other organisations have produced codes of practice and other documentation on this issue, discussion has appeared in magazine and newsletter articles, and to a lesser extent in books and journals from across the literature of computing, law, politics, information management and librarianship, science and technology, not to mention the popular newspaper and magazine press. Some of these writings can justly be described as contributions to a moral panic. It is therefore important that an organization like the Council of Europe, should form soundly-based views on the subject. The 5th European Ministerial Conference on Mass Media Policy, held at Thessaloniki, Greece, 11-12 December, 1997, responded to the commitment of the Second Summit of the Council of Europe held at Strasbourg, 10-11 October 1997 to seeking common responses to the development of the new information technologies, by taking as its theme 'The Information Society: a challenge for Europe'. The Conference produced a declaration and two resolutions with many implications for communications and information networks. These documents stress self-regulation by information providers and operators, and public education in the new technologies. This is emphasis is coupled with concern about the misuse of technology by those who promote violence and intolerance and show no respect for human dignity. The idea of 'universal community service' is invoked as a guiding principle for the resolutions, and this idea could also be taken as an implicit principle behind the content of the present report. The report is written from the perspective that every citizen needs to have free access to information and the ability to express opinion as freely as possible, so that society itself can react dynamically to change and continue to mature. It is also the writer's opinion that this applies almost as strongly to children as it does to adults. They need to be able to find answers to the questions that trouble them, and must at some stage in their development be allowed to judge for themselves what knowledge they need. That some people should wish to avoid unnecessary exposure to certain forms of expression on their own behalf, and on that of their children, is, however, natural. This is not a justification for the general suppression of any category of expression. It is a reason why it should be accepted as a legitimate exercise of choice for individuals, families and organizations like schools, youth groups, and even libraries, which take on some of the role of families, to seek methods to limit their own exposure to certain forms of expression. 1.1 Aims and objectives The purpose of this report is:
1.2 Definitions There may not be complete mutual understanding of terminology between those who work with information through media like the Internet and those whose main concerns are ethical and human rights issues. Therefore a slightly more detailed description of what is meant by 'networks' (the Internet in particular) for the purposes of this report, and an indication of the sense in which terms like 'freedom of expression' and 'censorship' are used, is provided here. 1.2.1 Networks Networks link computers, within a single site (local area network, or LAN) or via telecommunications systems to link geographically dispersed sites (wide area network, or WAN). Such networks may provide users with access to shared facilities (files, software, databases, printers, fax machines), but may also provide access to communications (email and conferencing). Networks can be linked through gateways to other networks. Where this uses the common protocol and addressing system tcp/ip, an internet is created. The global, public access network, known as The Internet, is the sum of all such networks that are in communication with one another. The significant thing about the Internet as a carrier of information, is that it has no centre and no controlling authority. It grows in an organic way as more networks are added. The early Internet had a primarily military function in the USA, and its distributed structure, both in terms of communication routes and information input, make it resistant to military attack or control. This robustness, combined with the international character of the Internet, also makes it difficult for civilian authorities to control. Matters concerning the Internet protocols and other network-wide issues, such as domain names and the IP addresses of networked computers, are dealt with by the Internet Society. The effectiveness of the whole system of communication is dependent on users accepting the guidance of the Society. The user obtains access to the Internet through some kind of access provider, typically an academic institution or a commercial organization dedicated to this form of service. The user's navigation of the Internet is facilitated by various providers who may offer some form of directory, sites which promote a range of information providers, or the search engines which compile information on sites so that the user can perform a keyword search for relevant content. Internet content is made available by content providers who vary from individual enthusiasts to the largest commercial organizations, many of them looking to the Internet as a source of profit. Whilst most content is in the form of text, there is an increasing quantity of more technically demanding content in forms such as audio, animation, or video-conferencing. Content is either accessed in real time, or downloaded on to the user's computer. Questions of intellectual property, liability for information content, and the means for monitoring and charging for commercial transactions are major issues associated with Internet content. Internet hosts, or servers, provide the computer capacity either for their owners' information, or only, as in the case of Usenet news servers, provide facilities for others. For most users certain sources of information and communication available via the Internet are more important than others. Newsgroups are highly significant as a means of public communication. Users post messages to one or more of the thousands of groups which deal with specific subject areas, often as contributions to a continuing thread of discussion. This is a comparatively ephemeral mode of communication, as material is generally removed after a time interval. Messages are, however, frequently re-posted, usually without the explicit consent of their originators, to one or more other groups. More permanent information public sources are generally on the World Wide Web. Web sites provide access to files in HTML (Hypertext Markup Language) and are connected by hypertext links. Electronic publishing via the Internet is comparatively easy for anyone with access to a server and the ability to work with HTML. Thus both informal communication and publishing on the Internet are democratically accessible to an unusual degree. This is, in turn, the source of the strength of the official and popular anxieties about the Internet as a means of expression. The Internet is, however, a means for purposive individual enquiry. Although accidental discovery of information does happen frequently on the Internet, this is not the same thing as the casual exposure to images and ideas that inevitably occurs with mass media. This likelihood of casual exposure is central to a rationale for the control of the content of mass media, but it is not applicable to networked information to anything like the same extent. Rather than concentrating on the dangers arising from some networked information, it would be better to stress the value of the access to information on a very wide range of topics, in detail, with little restriction, which the Internet offers. There are risks associated with false, misleading, inappropriate, or, what may be judged by some as, harmful information which is available on the Internet. These have to be accepted as a reverse side to the benefits of freedom of expression in society generally, and the Internet is no exception to this. 1.2.2 Freedom of Expression and Censorship. Freedom of expression is effectively defined in Article 10 of the European Convention on Human Rights, in other international conventions, and in the constitutions and legislation of various countries. Most notable is the USA, whose Constitution, in its First Amendment, prohibits Congress from 'abridging the freedom of speech, or of the press'. The protection of personal privacy, which is provided for by the European Convention in its Article 8, is also highly significant, particularly in the development and articulation of unorthodox and unofficial viewpoints. It is quite clear that the declarations in favour of freedom of expression are intended to protect both popular and unpopular ideas, but certain limitations on the scope of the freedoms offered are almost invariably specified. Thus Article 10 of the European Convention refers to national security, prevention of disorder and crime, protection of health and morals, etc. Exceptions to the universality of the principle can all be supported by strong arguments, but it needs to be remembered that these are also precisely the arguments brought forward to justify systems of official censorship. In the context of libraries, the American Library Association (ALA Intellectual Freedom Committee, 1986) has referred to censorship as:
Official censorship can be a formal manifestation of the widespread popular fears of freedom of expression which do exist. These fears emerge strongly in the context of particular subjects (child pornography, for example), or particular modes of communication (the Internet, for instance). Prompted by such fears, groups associated with a particular religion, lifestyle, philosophy, economic interest, or political tendency, can seek to persuade legislators, courts of law, or the public at large to suppress a certain form of expression. This report will draw attention to numbers of such organizations, lobbying and pressure groups, and putative censors. It will also discuss the work of organizations supporting freedom of expression. It should also be remembered that there is a range of non-censorship constraints on the freedom of expression and the freedom of information. Whilst in a mature, democratic society such constraints may not have the overwhelming influence that they exercise in less developed countries, they are still present. Prominent amongst such constraints are intellectual property rights and other mechanisms which arise from the ownership of information, the ownership of the media through which information is disseminated, and the associated costs to the user of access to information. The present easy and comparatively inexpensive access to the Internet, and freedom to use it, may be looked back on as a golden age from the vantage point of a future more commercially-constrained communication environment. This type of issue related to access presents too large and complex a group of topics to be mentioned more than tangentially in this report. 1.3 Methods The report is based on a sampling of current sources of opinion and news. This was carried out through:
1.4 Expected outcomes The intention was to produce a report which contained:
2. Debates on Freedom and Control 2.1 History of the debate Networks are widely perceived as being largely in the hands of their users, especially since the Internet itself has no central control. In this, networks resemble the printing press, which, because access to it is entirely decentralized, has always been a medium for dissent. Networks are, however, notably different from the broadcast media which, because they rely on a central means of transmission, have usually been successfully subjected to some level or other of regulation. The fact that the majority of the population, even in the industrialized countries, either only knows the Internet by repute, or has a limited personal acquaintance with it, means that fears and panics related to it can develop easily. Many of the journalists and other writers who comment on it, and the law enforcement officers whose responsibilities include matters of communication, also reveal a very limited knowledge and understanding of the Internet. This has made the short history of the Internet a fairly turbulent one. Fears which have been expressed include:
Such fears have found expression in some notable public cases, of which one or two examples will give the flavour. In the late 1980s the US Secret Service was regularly monitoring electronic bulletin boards for communications which related to the activities of hacker groups, believed to intend largescale sabotage of telecommunications networks. (Sterling, 1992) A false bulletin board had also been created in the hope of obtaining incriminating messages about suspected illegal activities. In connection with this, a raid (code-named Sun Devil) against suspected hackers had been organized, during which 28 premises were entered and more than 40 computers and 23,000 disks impounded. A company, Steve Jackson Games, which appeared to have only the most distant connection with the alleged offences, was investigated and a game that they were developing was claimed by the investigators to be a 'manual for computer crime'. The company and its employees then had to become involved in long and expensive efforts to obtain the return of their property, confiscated on suspicion. In 1991 a newsgroup was set up under the title alt.religion.scientology for the purpose of discussing this cult or religion. Members objected to this and from December 1994 onwards began to take various hostile actions against the newsgroup. These included claiming that messages posted to the group contained material which was copyright, and eventually trying to close the group down on these same grounds. It also took court action against individuals involved in the criticism of scientology via the newsgroup. Whilst the scientologists have not had marked success in the courts with their complex series of legal actions, these have indeed threatened to inhibit free discussion and criticism. In fact the effect has been more or less the opposite, with discussion and quotation from the documents of scientology multiplying and spreading because of the notoriety of the case. (Wallace and Mangan, 1996) In July 1993 two Californian citizens, Robert and Carleen Thomas, who supplied pornographic material via their 'Amateur Action Bulletin Board System' were investigated by Tennessee-based post office officials. They were then charged with a number of offences in the latter state, mainly involving the interstate transport of files containing obscene material. The case, and the conviction of the accused raised a number of issues including whether the material was actually obscene, whether it was likely to be seen by unwilling viewers, and whether when made available via a network it was actually 'sent' to those who obtained it. (Wallace and Mangan, 1996) The general trend of these and other cases was that communication on networks by certain individuals, or small and otherwise insignificant groups, attracted the attention of state agencies, or other powerful bodies. A variety of enforcement measures, based on different legal enactments was then taken, resulting in alleged injustices which aroused the interest of organizations set up to defend free expression. The Steve Jackson Games case drew together a group of issues connected with free speech, allegedly 'dangerous' information and computer crime, and Computer Professionals for Social Responsibility used the US Freedom of Information Act to obtain details of the Secret Service activities. Although the accused in the Thomas case were quite open about making their living from pornography, their case was taken up by bodies including the Electronic Frontier Foundation (EFF) and the American Civil Liberties Union (ACLU). This was because of the implication of the convictions that Internet communications that were legal when made in one jurisdiction could be found illegal when received in another. In the Church of Scientology case, the EFF intervened to suggest that litigation should be avoided because of the adverse effect on small providers. Probably the issue which produced the most public polarization of views arose from a study of pornography available on the Internet by Martin Rimm, of the Department of Electrical Engineering Department at Carnegie Mellon University. In 1994 he claimed that he had identified over 900,000 images with sexual content available over the network during a short period. He drew this to the attention of the University administration, which reacted by denying access via the University's computers to areas of the system, including Usenet groups designated alt.sex. This was opposed by the ACLU and other freedom of expression pressure groups, on grounds of principle, and because it also incidentally denied access to much valuable and entirely legal material (for instance discussions of safe sex). There was also disquiet about the effect on other institutions because of Carnegie Mellon's pivotal position in cooperative academic computer activities. (Faucette, 1995) Despite some publicly expressed doubts about Rimm's figures, they and the University's response have entered public debate as benchmarks of a kind. (Hoffman and Novak, 1995) The press has frequently published inflammatory stories about the alleged dangers of the Internet, often citing cases such as that at Carnegie Mellon. Again a few examples must suffice. Time magazine carried a cover story by Philip Elmer-DeWitt based on Rimm's findings, which has been cited as a key document in arousing public anxiety in the USA. (Elmer-DeWitt, 1995) The Observer (UK) of 25th August 1996 named people who it called 'pedlars of child abuse' over the Internet. It damned the defence offered by providers that they were avoiding 'unacceptable censorship', saying that,
The Daily Mail (UK) 6th Oct 1997 alleged that the Internet undermined nation sovereignty and presented a threat to commerce, morals, peace, law enforcement and social cohesion. The article argued that,
The volume of comment has been high. There has been some action against alleged offenders by law enforcement bodies, or by those who consider themselves in some way harmed by activities over the Internet. The sentiment that 'something should be done about it' has often been heard. 2.2 The US Communications Decency Act The Communications Decency Act (CDA) which was introduced in 1995, as part of a series of telecommunications reforms, was an explicit reaction to the anxieties which were being expressed in the press, and by various pressure groups. The measure, sponsored by Senators Exon and Gorton, effectively criminalized obscene material sent through electronic networks and provided for fines of up to $10,000 and imprisonment of up to two years. Specifically, it prohibited knowingly using 'an interactive computer service' either to 'send', or to 'display', any 'patently offensive' material to a person under 18. It was passed on February 1st 1996 and signed into law by President Clinton on February 8th. On the same day a group of organizations led by the American Civil Liberties Union (ACLU), called the Citizens Internet Empowerment Coalition (CIEC), filed a lawsuit challenging its constitutionality. Their detailed case against the CDA is summarized by Lappin (1996). The outcome of the legal actions was the decision of the Supreme Court known as Reno v ACLU which, on June 26th 1997, ruled the CDA unconstitutional under the provisions of the First Amendment to the US Constitution, arguing that:
Whilst the CDA was under debate, and the subsequent legal actions were in process, the Act focused discussion of the regulation of networks not only in the USA but throughout the world. After its defeat, President Clinton immediately called the White House Internet Decency Summit of July 16th 1997. At this meeting the emphasis was shifted to forms of regulation such as labelling and rating, filtering and blocking systems. There has already been an attempt to reinforce this approach by legislation, when Senator McDade of Pennsylvania introduced a Family Friendly Internet Access Bill, which requires Internet Service Providers to provide content screening software free, or at cost. This post-CDA phase of the debate may seem to represent the triumph of the free expression lobby, but the new emphasis on labelling, filtering, and blocking of sites has been described as more insidious and as dangerous. (ACLU, 1997; Lessig, 1997) 2.3 Key players The debate involves governments, law enforcement agencies, various libertarian political interests (including advocates of freedom of expression), churches, morality lobbyists, media, telecommunications and computer industry corporations and their representative bodies, librarians and other information professionals. As far as governments are concerned, a desire to deal with the dangers of information on networks is not confined to the USA. On July 30th 1996, the G7 Group of the industrialized nations and Russia agreed the principle of controls on the Internet including action against encryption that governments cannot break. Other political opinion may divide in unexpected ways. A conference on 'Cyberspace and the American dream' held at Aspen Colorado in August 1995, was promoted by the Progress and Freedom Foundation, a thinktank closely associated with US Speaker Newt Gingrich and the rightwing libertarian tendency in the Republican Party. (Right turn, 1995) Speakers at the conference included members of the Electronic Freedom Foundation (EFF), previously associated with President Clinton's Democratic Party government, but disillusioned by official policy on a range of freedom of expression issues. The analysis and conclusions of those present was overwhelmingly anti-regulatory and anti-Government. Network regulation is strongly advocated by religious and family-oriented groups which express concern over the material which might be accessed by children. For instance, the reaction of one such, the Family Research Council, to the Reno v ACLU ruling was that:
There are also feminist groups which strongly support Internet regulation. Because they feel that a free speech orthodoxy dominates the Internet, they call for strict regulation of a medium they feel can lead to an escalation of violations of women's rights. It is also true that there are feminists who oppose this line on the grounds that creating or endorsing powers to censor one kind of expression 'pornography' also hands the power to censor feminist speech to the same official sources, which they argue are dominated by sexist and anti-feminist opinion. (Carol, 1996) The anti-censorship lobby is represented by a number of loosely allied pressure groups, some of which are concerned with free expression as one of the human rights which they defend, and others which are specifically concerned with information and communication issues. The ACLU is an example of the former and the EFF of the latter. The ACLU was founded in 1920 to protect and expand American constitutional rights and civil liberties. It has a distinguished record in many important campaigns and court cases. Another example of such an organization with broader concerns which involves itself deeply with Internet freedom is People for the American Way, which was set up in 1980 to monitor and counter the agenda of the Religious Right political movement. The EFF, which has taken an important role in asserting the case for an unregulated Internet, describes itself as 'a non-profit civil liberties public interest organization working to protect freedom of expression, privacy, and access to online resources and information'. Its line is thoroughly libertarian and its statements often represent the thinking of John Perry Barlow, the author of 'A declaration of the independence of cyberspace', drawn up on 8th Feb, 1996, the date President Clinton signed the CDA into law. (Dority, 1996) EFF has local chapters in various parts of the world, and has launched the Blue Ribbon campaign, which encourages sites favouring freedom of expression to carry the blue ribbon motif. An allied organization, the Global Internet Liberty Campaign (GILC) is more explicitly international and links 38 member organizations, principally in the USA and Europe. The lawsuit against the CDA drew together a group which consisted both of such bodies, and also interested commercial organizations. The list is instructive:
Closer links can be seen in some cases. The EFF shares computer space, with Computers and Academic Freedom, which maintains an archive on information access in the academic environment. The Center for Democracy and Technology is concerned with public policies that advance constitutional civil liberties and democratic values in new computer and communications technologies. It shares an address in Washington DC with the Citizens Internet Empowerment Coalition. The CIEC, which was set up in February 1996 specifically to fight the CDA, describes itself as a broad coalition of library and civil liberties groups, online service providers, newspaper, book, magazine, and recording industry associations, and over 56,000 individual members.
3. Controversial Content on Networks 3.1 Sexual content It is apparent from any analysis of the searches made on the Internet that materials with sexual content are the most commonly sought. 'Sex' itself is the most common search term, and various permutations of 'pornography', 'erotica' , the names of magazines such as Playboy and Penthouse, and related terms predominate. (Markkula Center, 1997) As early as 1994, three of the most used Usenet groups were sexual in content. (Faucette, 1995) It is indisputable that there are both text and images available via the Internet which are sufficiently specific to satisfy any imaginable vagary of human sexual curiosity. That this is the case should be no surprise at all. One recent estimate of the annual turnover of the American sex industry (including printed, video and other materials) puts it at $8 billion. Some of this turnover represents illegal activities, but much takes place quite legally through shops, bookstalls, cinemas and mail order companies. It should also be noted that although American television largely excludes very explicit sexual content, pornographic material is an increasingly large element of the transmissions of television stations in countries including Germany and France. What happens on the Internet is therefore no different in principle from what takes place in the other media of communication. Sexual content is, however, the chief issue raised in the campaign to control communication on the Internet. Groups such as the National Campaign to Combat Internet Pornography, or People Against Pornography,(see list of Web sites) concentrate on this issue, which was the focus of the CDA's provisions. Their chief anxiety concerns child pornography, but feminist groups also express strong objection to pornography on the grounds that it is a means to demean and control women. Contributions to the Conference on Policing the Internet in London in February, 1997, by Gerstendorfer, Hughes, Kelly, Butterworth, and Muhonen developed this theme. (Policing the Internet, 1997) The core of the case for control of offensive material on the Internet is that whilst such material is bad in itself, it is particularly insidious on the Internet because of the way that it enters homes, schools and libraries. It is also alleged that it is possible for users, including children, to meet explicit sexual content inadvertently. It is certainly true that the major search engines will without difficulty retrieve sexual content for the searcher who asks for it. The extent to which the more explicit images, which arouse the most strong objection, can be found and viewed by the incautious or merely curious user is more open to debate. (Faucette, 1995) It may, indeed, seem like a rather convenient argument for those who fear freedom to justify restriction. 3.2 Hate speech The Internet is undoubtedly used for outpourings of abusive and threatening words, which are often directed at a particular race, religion, or sexual orientation. For example, anti-Jewish sentiments are common and virulent. Materials from the Institute for Historical Review (IHR), a California based organization devoted to the denial of the Nazi Holocaust which resulted in the deaths of millions of Jews and other minorities, are widely available through various Web sites. There are also Usenet groups which focus on this theme, with one (alt.revisionism) wholly devoted to it. Abusive terminology is habitually used to refer to Jewish people in exchanges in such groups, and versions of the blood libel appear, often in forms familiar from the fictitious document 'The Protocols of the Elders of Zion' which has circulated in print for decades. (Capitanchik and Whine, 1996) Other forms of racism, misogyny, anti-gay, anti-religious and similar abuse can also easily be encountered on the Internet. Whilst this is cause for general concern, it has been more addressed by organizations representing the abused groups than by the campaigners for control of the Internet. 3.3 Heterodox politics Use of the Internet to voice hate speech is closely related to its use for political discussion and organization by fringe and extremist groups. Neo-Nazi groups communicate via bulletin boards and newsgroups such as the Resistance Bulletin Board Service, and in the USA the Liberty Lobby, a racist umbrella group, sponsors the Logoplex BBS. (Capitanchik and Whine, 1996) The American religious right makes prolific use of the Internet, and the home page of Pat Buchanan (former US presidential candidate) containing his thoughts on family, faith and freedom, can be reached by links provided by Web sites on gun control, white supremacy, abortion and the other key themes of far right politics. (Newey, 1996) The Euskal Herria Journal, an online journal supporting Basque independence, has been accused of favouring the terrorist organization ETA. (Watson, 1997) Anarchist groups are enthusiastic users of the Internet, relishing its distributed and uncontrolled structure as a matter of conviction. (Atton, 1996) Examples of controversial political content could be multiplied. The effectiveness of political use of the Internet has been argued in relation to the disturbances in Serbia, 1996-7. (Bennahum, 1997) Described as 'The Internet Revolution' this campaign of political action and civil disobedience had to contend with strict control of the established communications media by the Milosevic government. Although there were only an estimated 10,000 possible Internet users in the country, their political significance was disproportionate to their number, most being students. They used the Internet in various ways. Demonstrations were called and coordinated through a service called SezamPro, even though it only had 22 dial-up lines and 3000 users. A radio station, Radio B92, which supported the opposition, was banned, but its broadcasts were re-routed via the Internet, using RealAudio. Overseas reporting of events was much improved by information communicated on the Internet, thus putting pressure on the regime. 3.4 Dangerous topics (Drugs, weapons, etc.) The Internet carries its share of information on topics relating to dangerous subjects and devices. There is, of course, much content relating to weaponry posted by gun lobby and similar groups in the USA. Shin Bet, the Israeli intelligence agency, is said to believe that encrypted instructions from Hamas, the militant Palestinian group, on terrorist attacks,
There is also publicly available information of an obviously dangerous kind available on the Internet. For instance, a bomb-making manual, The big book of mischief - the terrorists' handbook, has been available via the newsgroup rec.pyrotechnics. (Capitanchik and Whine, 1996) A Web site, provided by someone using the alias Candyman, is a collection of information on topics such as drugs, phone phreaking, techniques for killing people with the bare hands, and bombs. His justification for this is couched in freedom of expression terms. He claims that,
Despite such assurances, these examples illustrate why there is anxiety. 3.5 Defamation The Internet, particularly Usenet groups, is known for the freedom with which opinions of individuals are expressed and the abusive language that is often used. The practice of 'flaming' consists of subjecting individuals who have given offence in some way to uninhibitedly abusive public messages. This can easily contain material which is libellous. Although there have been only a small number of cases of libel actions resulting from statements communicated via the Internet, this certainly fails to measure the volume of libellous content. Anyone sufficiently determined to distribute a 'defamatory' statement across the system can deter prosecution by the sheer ubiquitousness of the material. For instance, McDonalds' took out a libel action in the UK against two people who had distributed leaflets criticizing the company. Whilst this was in progress, the original statements and much other arguably defamatory material were placed on a Web site called McSpotlight. This was held on servers in Holland, Australia, New Zealand and the USA, with the potential to distribute it even more widely if the need arose. (Katz, 1997) In another example, a poem by James Kirkup, which had been condemned in the British courts as a 'blasphemous libel' (it was a homosexual meditation on the crucifixion of Christ) was found on a US Web site. The UK Crown Prosecution Service attempted to take action against the Lesbian and Gay Christian Movement on the grounds that their Web site contained a link to the offending American site. The Prosecutor was unable to sustain this action and the case was dropped. 3.6 Official Secrets Similarly, the ability of the Internet to make official secrets public exceeds that of the media previously available. Once posted, a message tends to be re-posted in other Usenet groups or Web sites. 'Mirror sites' in countries other than that of an original posting are frequently used to evade attempts to suppress a particular item. The potential of this was exploited by Richard Tomlinson, a former member of the British secret service agency MI6, who wished to publish a book telling of his experiences with the agency. When the British Official Secrets Act was invoked to prevent him, he threatened to release the text on the Internet from a secret computer source where it was held in readiness for distribution. (Katz, 1997) In another case, Nottinghamshire County Council (UK)'s Joint Enquiry Team Report (JET) into the 1988 Broxtowe 'Satanic Ritual Abuse' case was not released for public attention. However, it was unofficially published on the Internet by a group that argued that its content was important public knowledge. Nottinghamshire took legal action for breach of copyright in June 1997, immediately after the report was posted. A total of 35 mirror sites were created around the world, forcing the Council to withdraw the action as hopeless. (UK Jet Report Controversy, 1997) 3.7 Privacy It must not be forgotten that networks carry a great deal of private communication as email, and in other forms such as computer conferencing. A proportion of the controversy over networked communication concerns freedom of private rather than public expression. 3.7.1 Surveillance Privacy of communication over networks is protected by the US law, and similar enactments in some other countries. In countries where the rule of law is precarious, such legislation is habitually infringed by law enforcement and intelligence agencies, and in countries where the law is formally respected, permission for 'wire-tapping' can be legally obtained if good cause is shown. The American National Security Agency (NSA) and Britain's GCHQ are alleged to intercept enormous numbers of international messages each year, ostensibly to prevent the use of networks for criminal purposes such as drug-trafficking, paedophilia and terrorism. Thus a senior British police office alleged that,
3.7.2 Encryption The quotation above was a reference to public key encryption, a system by which the potential receiver of messages makes public a key with which the messages can be encrypted. The receiver holds a private key which is the only means by which a message encrypted by the public key can be decoded. The use of such a method is a natural response to the real or imagined sense that surveillance is likely to break the secrecy of some message or other. Secure public key encryption for ordinary users is obtainable through the use of a freely available program called Pretty Good Privacy (PGP). It is significant, however, that its creator Phil Zimmermann was arrested and prosecuted by the FBI on the grounds that his program would be available to America's enemies, who might use it for some such purpose as espionage. To persuade or oblige network users to trade part of the security PGP gives to them, the US government has offered alternative, and even more secure, means of encryption. The US Congress is currently debating a Security and Freedom through Encryption measure, known as SAFE. This would impose controls on the manufacture and use of encryption in the USA. No encryption product could be marketed unless it contained a feature that allowed immediate decryption of a user's messages, without the user's knowledge. (New FBI Draft, 1997) The British government (before the change of control in May 1997) issued a Consultation Paper on 'Licensing of Trusted Third Parties for the Provision of Encryption Services' which followed a broadly similar line. (First Report, 1997)
Legislation is the first of three main types of response to the problems presented by controversial content of the kinds outlined in Section 3. The others are filtering and labelling systems (Section 5), and ethical approaches (Section 6). The failure of the US CDA has re-directed attention to the other approaches. This does not mean that the law is irrelevant, or that people do not still look for solutions through the making of new laws. 4.1 Relevance of existing law John Perry Barlow's claim that cyberspace is independent and that existing governments have no right to rule there, expresses a common sentiment, even a fear. As he put it,
This is, however, simply a romantic vision of cyberspace. The British lawyer Graham Smith suggests a harder reality:
It is true that in many cases the appropriate jurisdiction, or jurisdictions, may be extremely difficult to identify. In an attempt to deal with such difficulties, the British Computer Misuse Act of 1990 explicitly claims jurisdiction over an alleged offence whether it was the accused, the computer, or the offence itself that was in Britain at the relevant time. None of the relevant laws which might be used in the many jurisdictions where the Internet is used have actually been called upon very often. Therefore there is only a comparatively small body of case law arising from court actions related to communication on the Internet. (Diamond and Bates, 1995) Major texts on the applicability of existing law to cyberspace are in print, and no doubt others are being written. (Cavazos and Morin, 1994; Smith, 1996) Furthermore, despite the failure of the CDA there is fresh legislative and regulatory activity. 4.2 Proposed Legislation and Regulatory Systems 4.2.1 US States The level of legislative activity in US States relating to networks is at such a high level that a complete survey is inappropriate here. A few examples will illustrate the nature of this activity in state legislatures:
Other states have also had measures which shared features with the examples listed here. So far, it seems not to have been possible to enforce the measures which have been passed by the state legislatures. Certainly some of them are subject to challenges in the courts on the same grounds that were used against the CDA. 4.2.2 European Union The Union has succeeded in keeping the essential issues in mind despite the intensity of feeling which has influenced legislative action elsewhere. Concern was voiced in the European Parliament in April 1997, with a call for European and international action to drive pornography, paedophilia, and racist material off the Internet. However, the Union has so far avoided issuing Directives, which would have the force of law in the member states, on this issue. The debate in April called for a common legal understanding across the Union's member states as to exactly what constitutes illegal content, so that it can be made the subject of prosecution wherever the provider is based. In a major policy document, the 'Communication on Harmful and Illegal Material on the Internet' issued in February 1996 the European Commission had already set out views on this. (European Commission, 1996) The Communication summarised the complexity of the legal situation in Europe and of the problems to be addressed. It made a clear statement of the importance attached to the economic role of Internet service providers and re-asserted that the responsibility for content lies with authors and content providers. Action against service providers was seen as jeopardizing the ability of users to obtain access to other, legally protected, content. The Communication suggests that:
This line of thinking was developed further in July 1997 at a European Ministerial Conference in Germany. The Conference issued the 'Bonn Declaration' which contained a clear statement on legal controls applying to the Internet. It emphasized the importance of clear definition of the various Internet players' differing legal responsibilities for content. In particular it was recognized that there is a need for a clear distinction between the responsibility of those who produce and circulate content, and the Internet Service Providers who act as intermediaries. Since then, the European Parliament, at its session on Oct 24th, has recommended testing of filtering and screening devices, and called for a programme to inform parents on how to protect children from harmful content. Since then, various aspects of this developing European Policy have been drawn together in a very comprehensive 'Action Plan on promoting safe use of the Internet'. (European Commission, 1997) This outlines the issues and the state of policy development within the Union, then sets out a plan with four action lines. These are:
The Action Plan was issued on November 26th 1997, and has to be considered by the European Parliament and the Council of Ministers, before it can be adopted. It is significant that the Thessaloniki Ministerial Conference of the Council of Europe (see Section 1) in December 1997 passed Resolutions and approved an Action Plan broadly complementary to The EU Action Plan. Whilst the thrust of the EU Action Plan was 'safe use of the Internet', that of the Council of Europe was directed towards 'freedom of expression and information'. The EU Action Plan shows a strong concern for commercial issues, particularly seeking to protect service providers. This leads to an emphasis on community safety, and a strong emphasis on the testing of filtering devices. The Thessaloniki Action Plan concentrates on promoting freedom of expression, whilst studying misuse of networks and exploring possible action against misuse. Both Plans are strongly directed towards self-regulation, the developing of public awareness, and support measures through work on legislation. Although the two plans approach matters such as filtering and new legislation from rather different directions, they both call for work on the same areas, and suggest nothing that could not reconciled with the other's approach. 4.2.3 Individual European countries European governments have approached the issue both through taking legal action, and by encouraging self-regulation. The German government felt itself obliged by its laws against the spreading of race hate, and against aspects of pornography, to take action against certain Usenet groups. In December 1995, at the behest of the Bavarian prosecutor's office, it ordered CompuServe to prevent access to various sources of illegal pornographic and race hate material. To comply, CompuServe had to temporarily suspend the relevant newsgoups worldwide. All of the groups, except a few whose content was very specifically identified as illegal, were subsequently reinstated. CompuServe began, at this time, to offer the use of CyberPatrol filtering software so that parents and schools could control access to content generally. The German government also tried to bar access to race hate material on Web sites from Canada, but backed down when alternative access was offered by a number of mirror sites in the USA. In response to these and other similar instances, the German Federal Government has drafted a 'Law to Regulate the Conditions for Information and Communication Services'. This asserts the liability of information service providers for material passing through their services. (Kuner, 1996) The Netherlands and Britain, in contrast, have favoured self-regulation (sometimes in a form better described as self-policing). In the Netherlands it has been felt necessary to try to limit the spreading of paedophile material, and the authorities have sought to do this within the provisions of existing legislation. In January 1996, NILP, an association of Dutch service providers was encouraged to set up a foundation to monitor offending material. The foundation seeks to persuade offending providers to remove illegal material, reporting the offence to the police if it continues. (Vitiello, 1997)This balance between continuing respect for free expression and the upholding of the law is more or less paralleled in Britain. The Internet Watch Foundation (originally Safety Net Foundation) was set up in September 1996 by the two leading associations of service providers ISPA and LINX. (See list of Web sites for URL)IWF pursues the three goals of rating, reporting and responsibility. The reporting element means encouraging the public to report offensive sites to them, so that they can request police action if necessary. The number of British-based sites reported has been small and there have been no prosecutions in the courts. As another element of its programme the foundation continues to press the case for ratings systems. (Watson, 1997) 4.2.4 Asian countries Asia provides examples of the whole spectrum of policies towards the Internet. Thailand, at one extreme, makes little restriction and has the highest levels of use in South East Asia. Burma, at the other extreme, completely outlaws Internet access. Restrictive policies are more common in Asia than liberal approaches. In November 1997 Vietnam granted licences to four organizations to sell Internet services from a single state-controlled provider. Before this stage was reached, the party politburo had hesitated because it feared both the introduction of information about 'unhealthy cultures and lifestyles' into Vietnam, and the theft of national secrets. Traffic on the Internet is to be monitored and filtered by government agencies and the exchange of encrypted information is outlawed. (Jellinek, 1997) Singapore has adopted an approach designed to reconcile its aim to make the country the most wired in the world with its authoritarian attitude towards dissent. Since 1996, service providers must be registered, users are legally responsible for the material they receive and send, and much content is blocked on political, moral, and religious grounds. The development of the Internet in Singapore is a kind of laboratory experiment in controlling information whilst trying to obtain the benefits of the widest access to information resources. (Ang and Nadarajan, 1996) In China, since 1996, all Internet users have had to register with the police in a complex and (by Chinese standards) expensive way. Additionally, the single service provider, China Internet, blocks political material, and there is a high level of surveillance of Internet activity accompanied by intervention when activity is deemed inappropriate. The number of people with computers, the money to afford access, and the command of English required to make effective of the Internet is small at present, so close control of their usage is comparatively feasible. This pattern of restricting access to Internet content is referred to as the 'firewall'. The longterm aim of policy seems to be to combine these restrictions with a Chinese alternative to the WWW, the China Wide Web, which will be a commercial network, but also subject to control. (Usdin, 1997; Barme and Ye, 1997)) 4.3 Role of law enforcement agencies Some of the examples of policy discussed above involve roles for the police and other law enforcement agencies in administering Internet access on a day to day basis. In other cases the police might only be involved when there is some specific complaint which is to be brought to the courts. However, between these two distinct approaches there can still be police involvement. The Clubs and Vice Unit of London's Metropolitan Police has been pro-active in encouraging Internet service providers to prevent the circulation of obscene material, particularly that with paedophiliac content. In August 1996, the Unit called a meeting for ISPs at New Scotland Yard. In the words of Superintendent Martin Jauch they were told that 'We thought they were breaking the law and it was time for the industry to address that problem.' (Policing the Internet, 1997 p.30) This approach places the ISPs under threat and seems intended to encourage them to take on the role of censor, when neither does the state itself take censorship powers, nor have the courts made rulings on whether material is illegal or not. Reportedly, a number of newsgroups were removed by ISPs as a result of this pressure, even though most of their content would be legal in Britain. (Rodrigues, 1997) This form of policing purports to be encouraging self-regulation, but it actually deputes a censorship role to commercial organizations. Even if this were, in principle, a reasonable thing to ask, the ISPs have neither the will nor the legal expertise to perform the role effectively. A somewhat different approach now seems to have replaced this form of indirect police intervention. This is the setting up of the Internet Watch Foundation, referred to in 4.2.3 above. Although this is constituted as a self-regulatory body, controlled by a management board drawn from the Internet industry, educational, consumer media and libertarian groups, it is still described by some as a permanent threat and inducement to censorship at the ISP level. What is more, police intervention still seems to continue. The Campaign for Internet Freedom UK, had its site shut down by the providers Easynet in September 1997, ostensibly at the behest of the 'Anti-terrorist branch' of the police. (Watson, 1997) The site is now viewable via an American provider (See list of URLs), but it is important to remember that it was closed not by direct government censorship, or by formal agreement within an industry self-regulatory body, but as the result of pressure placed on a service provider.
5.1 The Metadata Dimension Filtering and labelling are discussed here essentially in the context of excluding access to material. However, they can also be regarded as part of the major movement to attach metadata to information on the Internet so as to enhance access. (Dempsey and Heery, forthcoming 1998) The World Wide Web Consortium (W3C), which is the main standards forum for the Web, has been working on an architecture to accommodate metadata. This Resource Description Framework (RDF) is intended to provide for the very wide range of different metadata needs experienced by those creating and using material on the Web. What is required of the RDF is that it should support a variety of resource description models devised to satisfy all these distinct needs. RDF will use XML, the markup language which is being developed as a successor to SGML, as what is described as its transfer syntax. This will make it possible to take advantage of the various tools which are being developed around XML. PICS, which will be discussed later in this chapter, is one example of a metadata activity taking place in this context. Another important illustration of these resource description models are those being worked on under the name of the Dublin Core. The Dublin Core is a metadata set with 15 elements, which is intended to enable searchers to discover electronic resources on the Web. It is intended to be: simple (much simpler than library systems such as AACR2) so that it can be used by non-cataloguers; interoperable between different disciplines which have their own differing standards for resource description; internationally acceptable; flexible, to the extent that it can be used in associated with more elaborately structured forms of description if required; and, finally, capable of coexisting with other metadata packages (which might relate to such matters as the privacy status of the resource, or terms and conditions for use of intellectual property). These characteristics are typical of what might be looked for in an effective metadata package and most of them could just as reasonably be looked for in a package, like PICS, designed to identify material which a user did not want to see, or did not want others to see. Metadata can be applied to resources in various ways. The most obvious way is to embed it in HTML documents themselves, using the <META> tag which is provided. The HTML 4.0 specification released in July 1997 permits a richer form of description than previously under the <META> tag. The data provided in this way is an integral part of the resource and is picked up by the agents which index a Web resource, along with all the resource's other elements. Alternatively an organization which collects and manages metadata records which are not embedded in the resources themselves (in just the way that libraries catalogue their holdings) can label resources for additional purposes. Such purposes could include rating of content according to suitability for different user groups. Finally, it is possible to envisage the management of metadata records by a single agency which would ensure that various descriptive models were interoperable to the extent that users could identify precisely the form and content of resource they required using a single query model. 5.2 Content filtering and recommendation systems Having set labelling in a metadata context, it is necessary to look at the filtering process which is made possible by the availability of suitable metadata. This is also a much broader activity than the mere filtering out of distasteful content (although this is the usual sense in which the term is used in the context of discussions of freedom of expression). The bewildering wealth of resources on the Internet, and the difficulty of assessing their relevance and quality, has led to explorations of filtering systems, using software developed for the purpose. Such software can use various means to identify resources on behalf of the user who installs it or takes advantage of services which provide assistance with the filtering of information. Early developments of this kind were often referred to as collaborative filtering systems. It is now more common to refer to recommender systems, as the term 'filtering' does have some connotation of exclusion, and not all such systems are truly 'collaborative'. (Recommender Systems, 1997) 5.2.1 Filtering for recommendations Recommender systems are, in spirit, the same as the published consumer reports which test and assess products and services such as restaurants and hotels. They pass on the assessment made by one or more individuals, expert or non-expert, for the benefit of others. Symbolic awards of stars or rosettes are often used to sum up the results of this process. The recommendations, whatever the form in which they are expressed, are a type of metadata. Electronic recommender systems usually work by aggregating inputs from interested people and directed the aggregated information to those who seek guidance. Their distinctive contribution may lie in the aggregation itself, or in their ability to make good matches between recommenders and those needing advice. Recommendations may be entered into the system explicitly, but some systems gather the evaluations implicit in recommenders's use of resources (in the form of references to URLs in Usenet postings, personal bookmark lists, or the amount of time users spend with a resource). The recommendations may be anonymous, pseudonymous, or tagged with the recommender's identity. The aggregation may take into account a system of voting weighted according to some scheme such as past agreement between recommenders, or personalized weighting. In some systems evaluations and content analysis may be combined to produce a recommendation. The recommendations can then be used in various ways, such as to filter out resources that are not recommended; to sort and present lists of resources according to numeric evaluations; or to present evaluations in a display with the items to which they refer. This may seem to apply best to the recommendation of Websites, but it is claimed that even Usenet news articles, which are very numerous and have short lifespans, have been effectively rated and recommended. A number of existing products and services provide recommendations using a mixture of the methods referred to above. They include GroupLens, Fab, ReferralWeb, PHOAKS, and Siteseer, all of which are described and analyzed in some detail in an issue of Communications of the ACM. (Recommender Systems, 1997) 5.2.2 Filtering to exclude content Having demonstrated that filtering is merely one application of metadata, and only one aspect of the filtering and recommendation function, it is, nevertheless, a central feature of the discussion of how to deal with offensive content on networks. Even before the US Supreme Court rejected the CDA in June 1997 filtering was discussed as an important alternative to government intervention by legislative action. It is now central to the debate. This approach has been described as 'the privatizing of censorship', (Lasica, 1997) and there are two main routes by which it can be achieved.
The means used to make this possible is the Platform for Internet Content Selection (PICS). PICS is an HTML standard which makes it possible to filter material on the Internet. (Resnick, 1997) It establishes a consistent way to express content ratings, which can then be attached to specific resources according any one of the systems which might be available. The user can then filter out any content which is revealed not to meet self-set criteria of acceptability. PICS is not itself directed against any category of material: it provides the means by which content rating schemes from the whole range of possible sources can be operated. The intention is that individuals, families, organizations, ISPs, or even nations, will be able to select the content rating systems they prefer, and use PICS to put them into operation. (Resnick, 1997) This is also the position of the European Commission, which discussed content rating systems, in 1996 and endorsed the value of a multiplicity of such systems in allowing users to select the one which reflects their values. (European Commission, 1996) The creators of browsers and search engines, including Netscape and Microsoft declared their agreement to provide PICS capabilities in their systems at the Fifth International WWW Conference in Paris in 1996. Also CompuServe expressed the intention to use PICS to put ratings on its content as it moves on to the Web, and to provide CyberPatrol software for its members's use. The swift acceptance of PICS by the industry as a whole has led to explorations of the extension of its use to cover code signing, privacy and intellectual property rights management. PICS's own Web pages (see list of URLs) provide the basic details, connect the reader to other related information and deal with a number of key FAQs. The filtering software which makes use of the ratings which PICS presents, so as to block access to certain resources, is a crucial element in this whole system. A number of products are named above. One of the first and best publicized is SurfWatch. It has its own Web page, which proclaims that 'it is a real alternative to Internet censorship, giving parents and educators the opportunity to limit unwanted material locally without restricting the access rights of other Internet users'. (Internet Censorship, 1997) There are two main objections to filtering software. First, the exact methods by which products identify material to be blocked may not be fully transparent. Second, products are often alleged to block valuable material implicated only by the crudity of these methods. For instance, Web sites for English pubs were blocked because of references to alcohol; a real estate site was blocked because it used the same Internet access provider as a pornography site; part of the US White House site was blocked because it referred to the Clintons and Gores as 'couples'. (Robot as Censor, 1997) What is clear is that large numbers of sites are blocked by many of the products. A list of the number of sites that were alleged to be blocked has been posted on a Web site (although it is possible that this information may not be reliable). The list suggested that, on various different count dates, CyberPatrol blocked 13,000 sites; SurfWatch 4,500; the Internet Filter 107; NetNanny 801; and CyberSitter 820. (Censorware Search Engine, 1997) An organization which supports filtering, Filtering Facts, has tested and (so far) recommends four products, Bess, CyberPatrol, SurfWatch, and Websense, as meeting its criteria, which are:
5.3 Rating and Labelling Systems Ratings or labels can be applied to resources either by some bureau designated for that purpose (a third party), or they can be applied voluntarily by the creators or the distributors of the resources. 5.3.1 Third Party Rating PICS's creators in W3C envisage that there will be a multiplicity of ratings services, covering shades of opinion from across the whole spectrum of human preference. Out of these, people can choose one which reflects their own values. Certainly it seems possible that all kinds of bodies with a religious, moral, political or ideological viewpoint will be interested in devising a scale or matrix for ratings. This is, however, a different issue from actually examining resources and deciding on a rating for each example. The likelihood that any but a major, industry- or government-funded agency could do more than apply ratings to a proportion of the most permanent and high-profile sites seems low. (Marshall, 1997) The expenditure of time and money required to rate tens of thousands of Web sites, let alone Usenet contributions, would not be available to small organizations. Existing systems which rate audio-visual products deal with much more manageable numbers, such as feature films for cinema distribution, videos for public sale or rental, and computer games and simulations. The use of third party rating is long established in the audio-visual industries. For instance, the British Board of Film Certification (BBFC), previously known as the British Board of Film Censors, has applied ratings to films which are to be shown to the public in cinemas for decades. The system has no legal force, but the local authorities which are responsible for the licensing of cinemas use it as their standard, only occasionally making specific decisions to permit the showing of an unrated film, or not allow the showing of a rated film. It is notionally enforced by the cinemas themselves. Films are viewed by the Board and awarded a distribution certificate essentially suggesting the age at which it is felt that they would be appropriate for young viewers. The certificate for a film indicates a suggested age, such 12 or 15, as appropriate for viewers, or designates it PG (Parental Guidance). In practice cinemas seem to regard the certification as being addressed more to parents than to themselves. One major rating system intended specifically for Internet use has been devised by the Recreational Software Advisory Council, a body which originally concerned with computer and video games. The system is called RSACi (with the 'i' standing for Internet) It rates resources on a scale of 0-4 on four content categories: violence, nudity, sex, and language. Thus a product with a rating with the structure 4 3 2 1 would be expected to contain: wanton and gratuitous violence; non-sexual frontal nudity; clothed sexual touching; and mild expletives. Alternatively the structure 0 1 2 3 would suggest: harmless conflict with some damage to objects; revealing attire, clothed sexual touching; and strong, vulgar language, obscene gestures, and racial epithets. 5.3.2 Voluntary self-rating If, as suggested earlier, extending rating by a third party right across the vast, and growing, number of Internet resources proves to be impossible, a system of self-rating offers a more practical possibility. In such a system, a site voluntarily compares itself with a set of criteria of the kind described above, and then indicates this rating on the site. Filtering software in the home could be adjusted to accept only sites which bore a rating, and amongst them only sites with certain levels of rating. The scale of activity this would demand is enormous. By October 1997 45,000 sites out of the millions available, had rated themselves according to some criteria such as those provided by RSACi. (Arthur, 1997) The implications of self-rating for free speech are also enormous. The holders of controversial opinions are either likely to be excluded from communicating with many people if they rate themselves honestly according to the crude and imprecise categories offered by ratings schemes, or discouraged from full and free expression by the need to show a bland rating for their material. (ACLU, 1997) Legislation to make rating mandatory has been discussed in the USA, and it seems a natural consequence of such a system. Certainly, without a penalty system to back it up, self-rating is likely to be subverted by some providers who would rate their sites in a deliberately deceptive manner. Within the Internet community there is no agreement on how to administer a self-rating system so as to avoid compulsion. There is also no clear idea for a means to ensure that all creators of sites, wherever they came from in the world, could be drawn into the system of self-rating. Unless they were, their sites would be filtered out by many users for no good reason other than being unrated. 5.4 Authentication of users A final element in systems of rating is what amounts to the rating of users themselves. The creator of a Web site or the moderator of a newsgroup could restrict access to all but authenticated users (effectively the members of a club, or registered customers). This would prevent access by children to material in sites only open to those with adult authentication.. (Cormack, 1997) The administration of this could be done in one of four main ways.
Whilst discussion of authentication is mainly a privacy issue, it will be seen from the above that it does have a logical place in the discussion of filtering access to resources. By placing responsibility for who accesses a site firmly in the hands of information providers, it answers some problems inherent in other means of dealing with the issue. What it does is to get providers to police access to their resources in rather the same way that the proprietors of 'sex shops' in Britain are required to limit entry to those over 18, and bars in many American states have to check driving licences to ensure that their customers are all over 21. There are websites, such as Adult Check, Adult Pass, Adult Sights, which offer verification services. Adult Check claims to be used by nearly 3000 websites which have 'adult' content. The way it works is that those who wish to have their status as adults authenticated for this purpose, respond to the site's request for details of name, address, and credit card. If these are accepted a password is issued for a payment of $12.95 per year. Even if the procedures used by those responsible for the site are sufficient to ensure that this information is accurate and consistent before the password is issued, it would be easy for a child to provide genuine details of an adult and then use this false identity. There is also evidence that the information held by these services is not secure, and that there are no means within the system to prevent one person using another's ID. (Markkula Center, 1997) 5.5 Research findings A good deal of commentary on filtering centres upon its obvious or alleged technical imperfections. Opponents of filtering regularly cite the imperfect way in which filters work as a key element of their case. This is usually based on fairly casual testing of the system. For instance, a small investigation at the Markkula Center, Santa Clara University, tested whether filters would block sexually-explicit sites, sex education sites, safe sex sites, sites with factual material on lesbianism, similar sites with gay information. It found that the filters it tested blocked all sites in the first category, permitted access to sex education sites, had mixed results with safe sex sites, allowed material on lesbianism through, but blocked gay sites. These interesting but inconclusive findings are reasonably typical of those mentioned in discussions of the topic. So far there has been little independent research which thoroughly explores the effectiveness of Internet filtering, and its implications for the use of the Internet as an information resource. An exception is the Internet Filter Assessment Project (TIFAP) which, although imperfect in a number of admitted ways, does provide a good deal of interesting information about the effects of filtering systems on the information seeker. This small project tested 18 examples of filtering software, looking at the way in which they blocked material by: keywords identified within their content; the names of particular sites ('intellectual de-selection' is what TIFAP calls this); and by protocols (all Usenet groups, IRC, or Telnet). Blocking sites was unsatisfactory because new sites spring up all the time, and until their content is identified as problematic they remain fully available. Blocking by protocols was revealed as extremely crude and over-inclusive. Blocking by keywords was tested by attempting to obtain answers from the Internet to over 100 questions typical of those asked in libraries, while the software was in operation. The questions explored the software's treatment of eleven topic areas,
The questions approached these areas directly (by asking for research material on controversial topics) and tangentially (by 'inadvertently' asking for terms which might be blocked, such as recipes for chicken breasts, or cultivation of rape seed). The testers found that the filters blocked information needed to answer questions in over 35 per cent of cases, reacting to keywords and blocking sites with necessary information on topics such as safe sex, organizations for gay teenagers, and arguments for and against the legalization of drugs. The significance of this was that all the filters were, to some extent at least, doing more than their intended role of preventing inadvertent or deliberate access to material that was offensive or indecent. They were actually hindering legitimate and useful information searching. 5.6 The Debate The debate on filtering has been carried on at a high rate of intensity ever since the defeat of the CDA seemed likely. As evidenced by the number of Websites representing organizations wholly or partially concerned with the issue, there is wide interest in the issue and a sharp polarization of opinion. The division seems to be between advocates of untrammelled free expression on the one side, and a spectrum of groups ranging from those pro-censorship to those anti-censorship, but favouring enhanced ability to avoid offensive material, on the other. A good deal of the debate centres on the use of filtering software in libraries on behalf of users, rather than on their use by families in the home. The ACLU and the American Library Association were not merely key actors in the campaign against the CDA, but they both oppose filtering very strongly. Their position is based on a respect for individuals' and families' right to make up their own minds what they look at, and a conviction that filtering hinders this. Other sources of argument against filtering are to be found in the Web sites of Censorware Search Engine, Cyber-Rights & Cyber Liberties (UK), Ethical Spectacle, MIT SAFE, and Peace Fire. (See list of Web sites for URLs). There have also been a number of press reports and magazine articles drawing attention to the possible implications of filtering, of which the following are a selection. (Kleiner, 1997; Lasica, 1997; Wallich, 1997; Watson, 1997; Winner, 1997) A thorough review of the issues and technology from this perspective is Cyber-Rights and Cyber-Liberties (UK)'s report 'Who watches the watchmen: Internet content rating systems and privatized censorship'. (Cyber-Rights, 1997) The basic technical case in favour of filtering is made very strongly by those responsible for PICS, particularly Paul Resnick of AT&T Labs, Chair of the PICS Working Group of W3C. (Resnick and Miller, 1996) Filtering is supported strongly by Filtering Facts, which is chiefly concerned with the use of filters in libraries. Its criteria for filterware selection listed in section 6.2.2 were devised largely for use by libraries. David Burt of Filtering Facts has countered a number of the arguments against filtering, including those arising from the admitted anomalies produced by the operation of products currently available. (Burt, 1997) Other sites which promote the case for filtering are Family Friendly Libraries, Enough is Enough, Library Watch, National Coalition for the Protection of Children and Families. (See list of Web sites for URLs) The form that the debate can take within the library community, between those who have chosen to filter or not to filter, is well illustrated by a published debate between the librarians of Austin (Tex) and Monroe County (Mich). (Branch and Conable, 1997)
The common tendency to portray the Internet as a kind of anarchy ignores a considerable will amongst many users to encourage responsible use. It also fails to take account of the way in which the public organizations providing Internet access, service providers and, not least, the library community, have developed policies for purposeful and considerate community and family access to Internet content. The quality of the work done by librarians on ethical alternatives to the regulation and policing of the Internet is widely recognized by informed commentators. For those who reject both the formal regulation of the Internet, and the use of technology to restrict access to content, these explorations of ethical approaches hold the best possibilities of convincing the general public and political leaders that the Internet is actually a force for good, rather than some kind of threat to the structure of society. 6.1 Netiquette and acceptable use policies 6.1.1 Netiquette The flooding of newsgroups, mailing lists or other elements of the network with messages publicizing the views of an individual, or advertising some product or service (spamming), and the sending of abusive mail (flaming) are one aspect of network use. Counterbalancing this is an evolving sense of the good manners which are required to make network use convenient and unthreatening for all. Codes of good manners for network users are referred to as 'netiquette' (sometimes nettiquette). Most of what they contain is unremarkable, amounting to a restatement of the elements of considerate conduct which apply in society generally. One widely-available codification of netiquette (Shea, 1994) is structured round 10 rules:
That such lists are more than just preaching by a few concerned users is evidenced by the numbers of outraged messages that appear in reaction to what are considered gross violations of good network behaviour. Intricate discussions within newsgroups of the rights and wrongs of some particular alleged infringement of the code are also not difficult to find. However, this concern with ethics is not the dominant tone of exchanges on the Internet. As one commentator points out,
The obvious implication is that there is a need for something more carefully worked out and widely accepted than the rather vague goodwill expressed in netiquette. 6.1.2 Acceptable use policies Whilst netiquette is essentially a guide for individual behaviour, acceptable use policies are the guidelines and requirements laid down by organizations relating to the conduct of users within their systems. The effort which has been put into creating such policies is considerable, and cumulatively they reveal a considerable consensus on what is reasonable use of the Internet. If all users followed the precepts of these policies, the perceived problems of Internet content would almost vanish overnight. Policies exist in enormous numbers. Many examples from schools, for instance, can be found on the Internet.A large number are listed in the Web site of Robbinsdale, Minnesota, School District. (See list of URLs) Advice is available for those compiling such documents, and one adviser discusses the issues under the following headings:
Other academic institutions, providers and networks also have their policies. The UK Joint Academic Network (JANET) set out a policy in April 1995 which effectively applies to all British academic users. In its ninth clause it spells out what is considered unacceptable use:
9.4 the transmission of materials such that this infringes the copyright of another person;
Internet service providers also have their policies and a collection of these is available online. (See list for URL of Acceptable Use Policies site). The policy of America Online (AOL) sets out its principles with helpful explanations of the various elements it covers, such as email, Usenet and chat. (See list for URL) The problem is obviously that whilst providers of access in the public sector (schools, universities, libraries, government departments) might be able to enforce their policies with members and employees, it is a more delicate matter for commercial providers to limit the activities of paying customers. 6.1.3 Industry codes of practice There is also work in progress to draw up industry-wide codes, which define more fully the obligations of the various types of provider. An organization called the Internet Content Register has, with cooperation from the UK Consumers Association, drawn up an Internet Code of Practice (ICOP). (Internet Code, 1997) This covers:
The organization offers certification for service providers who undertake to operate under ICOP. This is to be backed up by monitoring and the investigation of complaints. An agreement is offered between Internet Content Register and service providers, which has clearly-stated mutual responsibilities. Whether this initiative proves to be successful or not, it offers ideas as to what shape industry self-regulation through codes of practice might take. 6.2 Library Policies One regular theme that occurs in documents about self regulation of the Internet is tribute to the usefulness of library policies in addressing the whole issue. For example, Computers and Academic Freedom makes available an article which discusses how principles developed in the library world can be applied to public and academic use of computers. (Kadie, 1994) The principles discussed in this chiefly come from documents prepared by individual libraries, and by the American Library Association (ALA). Other important library organizations do not have such clearly articulated positions. The International Federation of Library Associations (IFLA) passed a resolution accepting the report of a Committee on Freedom of Access to Information and Freedom of Expression (CAIFE) at its meeting in Copenhagen in September 1997. However the recommendations in this document were of the most general kind and said nothing specific about networks. Subsequently, however, a permanent Committee of IFLA was established as an ongoing mechanism to address this issue. The Library Association (UK), whose members and secretariat do monitor and discuss freedom of expression on networks, also has only the most general policy documents, such as a Statement on Censorship (last revised in 1989). The ALA, in contrast, has consistently taken a strong stand in favour of freedom of expression and accepts that it must therefore show how freedom can be used safely and responsibly. The basic position of the ALA is set out in its 'Library Bill of Rights', as revised in 1980, taken in conjunction with resolutions on 'Access to Electronic Information' of 1996, and 'Use of Filtering Software in Libraries' of July 1997. (See ALA Web site, URL in list) These documents show a firm stand on US constitutional freedoms, which are held to apply to all, regardless of their age (subject only to parental guidance for children). The Internet is treated as analogous to a library, and therefore the principles operating in libraries are argued to extend naturally to Internet access on library premises. (Markkula Center, 1997) Filtering is accepted by some librarians, who draw a further analogy with book selection and the mediated access to parts of library collections which is normal for one reason or another. The crudity of the filters available has, however, created real difficulties for many librarians in reconciling their fundamental faith in freedom of information with the use of any filters. (Mason, 1997)In holding to policies favouring unhindered freedom of access, libraries are subject to pressure from organizations such as Enough is Enough. This organization provides a kit of questions and arguments for use by those wishing to encourage libraries to use filters to shelter children from offensive content. (Safeguards, 1997) 6.3 Personal and parental responsibility Finally, there are organizations which point out that ultimately individuals must accept personal responsibility for what they choose to view on the Internet. Furthermore, they must accept that they have responsibility for what their children view. This line of argument turns attention towards rules for personal and family responsibility, and towards ways of making positive selection of high quality material. This is the implication of the policies of an organization called The Children's Partnership, which in turn cooperates with other US organizations, the National PTA and The National Urban League. The Children's Partnership focuses on parental involvement, the quality of information literacy skills teaching in schools, and equality of access in the information society, rather than on negative aspects of Internet use. (Parents, 1996) As far as parental involvement is concerned, there are sets of suggested rules both for parents and children. The National Center for Missing and Exploited Children (see URL for Child Safety on the Information Highway) suggests the following seven guidelines for parents:
A similar set of rules which parents can encourage children to follow is offered by Pippin (see URL), an organization promoting positive use of the Internet:
The ALA, similarly, advises parents that the best way to ensure their child's safety on the Internet is to be there with them and to agree rules. (Librarian's Guide, 1997) Like several other organizations, it complements this by actively recommending sites which it feels will entertain and inform children. An original list of 50 has been expanded to more than 700, selected according to criteria that are displayed on the site. (Great Sites, 1997) There are genuine causes for concern arising from the nature of some material available over the Internet. However, these differ from concerns over print and broadcast material only because of the comparatively unregulated way in which networked information enters the home, library and educational environment. Public and official anxiety has expressed itself in three main approaches to the issue: legislative solutions, filtering and blocking, and self-regulation of the Internet and its use. Solutions based upon new legislation are generally inappropriate for various reasons:
Filtering and blocking is not a fully accepted solution on grounds both of principle and practicality. Objections in principle are more or less insurmountable, but some minor element of filtering could be used by parents, and those with parental responsibility, without deviating from the social norms which limit access by children to material in other forms. Use of filtering and blocking must be by choice and, if possible, consent. Their imposition without consent or the scope for challenge is surely just as unacceptable as any form of censorship. This is a compromise, but compromises are a normal part of the fabric of life. At present, compromises on filtering and blocking are difficult because the filtering process is crude in its effects, and there is a lack of transparency as to what is blocked, and why. Any acceptance of filtering and blocking, by those who favour freedom of expression, will require much improved technology, and high quality systems for the rating and labelling of content. An acceptable platform for ratings metadata is already available in the form of PICS which, despite hostility to it from advocates of complete freedom of expression, is not, in itself, automatically an instrument of censorship. It is the systems of rating which PICS would support which are more capable of being a means of censorship. Ratings systems require good sets of standards, and existing ratings standards, mainly designed to deal with recreational materials, are inadequately specific. To attach ratings metadata to content is ideally the work of content providers themselves, but good systems of third party provision are also acceptable. Properly funded ratings agencies, which enjoy the confidence of the Internet industry and users, would be needed if this alternative is to be pursued seriously. Self-regulation of networks and their content through codes of practice for content and service providers and users is, in principle, and for technical reasons, the best means of ensuring confidence in networks as communication media. The improvement and promotion of such codes of practice will be required if they are to capable of effective application by the groups at which their provisions are directed. Self-regulation should not mean self-policing, so the relative spheres of law enforcement agencies and regulatory bodies need to be clearly delineated. Finally, self-regulation needs to be supported by programmes to encourage responsible and informed use of networks, along the lines of those provided by the ALA and other organizations. Part Two: Access to communications networks, particularly the Internet, is an issue for the individual and for governments, but it is also an issue for those who have managerial responsibility for public access points. It is clearly the responsibility of individuals to make decisions on their own behalf about what they look at and read. At the same time, democratic governments have a responsibility to ensure that the information available conforms to the requirements of the law, particularly on matters such as national security, the prevention of disorder and crime, and the protection of health and morals. But at a public access point, the essential responsibility of those in charge is to give individuals the fullest possible freedom to seek for knowledge. This is a difficult role for two reasons. One is the availability via networks of content which is illegal in the jurisdiction where it is accessed (though not necessarily illegal where it was originated) and other material which, whilst it is not illegal, is widely believed to be harmful. The second is that networks are extensively used by young people for whom their parents still have legal responsibility. Popular opinion is that these young people are in danger of harm from certain kinds of content. The manager of a public access point must find ways to cope with this confrontation between the basic principle on which public access is provided, and the anxieties about access that the nature of some content creates in the public mind. That there is genuinely felt concern, rather than something created by the news media to increase interest in their publications and broadcasts, can easily be illustrated from some recent surveys of opinion in Britain. In the first Which? Online survey by the research agency MORI (Annual Internet Survey, 1998), 58% of respondents agreed with the suggestion that the Internet undermines the morality of the nation by making pornography and other illegal materials freely available, and only 13% said they would feel comfortable letting their children use the Internet unsupervised. At the same time, another MORI survey of children's attitudes suggested that parents might be right if they felt their children were potentially beyond their control in this area, since 30% of children were confident that they knew more than their teachers about the Internet (Children's attitudes, 1998). Finally, a slightly earlier survey of teachers' perceptions of what were termed 'the new entertainment technologies' revealed that teachers were seething with anxieties. Approximately 75% were confident that there was a cause and effect relationship between frequent use of computerised entertainment and various aspects of impaired learning (Miller, 1994). These concerns may, or may not, have a basis in reality, but they do create a particular environment of opinion in which public access points must be managed. This report uses the same methods as those which were employed in Part one of this document. It presents an argument relating specifically to network public access points, based partially on the material presented in part one, but also on a fresh sampling of news and opinion from the first six months of 1998. The intention is to contribute to the preparation of Recommendations by the Council of Europe, which will be of assistance to governments contemplating legislation or other actions which will have implications for the management of public access points, and also for those who are directly responsible for the day-to-day management of such points. 2. Network Public Access Points Although many families in Europe have a computer in the home (38% in Britain, with a higher percentage for homes where there are school age children) (Nation divided, 1997), this still leaves a large majority of the population dependent on public access points if they wish to make use of networked communication and information resources. Provision of public access is not at all evenly distributed between nations, nor it is it evenly distributed within nations. The highest levels of public access are available in educational establishments. Libraries and other information institutions, cultural institutions, and information kiosks, also give access. The extent of public access outside the educational sector varies considerably between individual institutions and types of institution. The importance of public access of this kind is that it is the source of any claim which networks might have to be a universally available phenomenon. Therefore the policies and administration of the hosting institutions are crucial to making freedom of expression and freedom of access to networked content matters of fact as well as principle. 2.1 Educational establishments Access to networks in universities and other institutions of higher education is more or less universal in European countries, and schools increasingly also offer access to both their staff and students. More than 85% of British secondary schools have Internet access, although the same is true of only 5% of primary schools (Blamire, 1998). As a range of government initiatives are in place to spread school access more widely in Britain, it is reasonable to expect something like universal access will be achieved early in the new millennium, and other European countries are also moving swiftly in this direction. Universities generally seek to have a networked computer in every staff office, and many are also beginning to offer network connections in the rooms of student residences. However, more commonly, access is obtained in a computer room, laboratory or library. Such facilities vary from a single networked machine in some schools through to literally hundreds in the learning centres of well-equipped universities. This has the implication that, for the most part, networked material which is accessed by teachers in schools, and by students at all levels of education, is viewed in public areas of their institution. 2.2 Libraries and other information institutions Internet access in libraries and information centres is an increasingly important form of public access. For the purposes of this report, access in academic and school libraries can be considered as an aspect of public access in educational institutions generally, rather than as part of library access. Research and other special libraries do, however, fit under this heading. Because of their closely defined and limited clientele, they differ from public libraries in matters of network public access. They, and access points provided in information offices specialising in tourism, education, employment, and other topics of importance to the citizen, have a defining clarity and singularity of purpose. The case in public libraries is clearly different, and it is public libraries in which the truly public access is beginning to be more common. Large numbers of British and Danish public libraries are substantially automated, and many offer Internet access. Universal connection of public libraries to the network is promised in Britain in the near future (New library, 1997). However, only 20% of German and Portuguese public libraries were considered as automated in 1996 (Thorhauge, 1997). Only 20 German public libraries had their own Website in 1997, but progress since has been swift with quite small public libraries, such as that of Zadar, Croatia, having put up professionally presented Websites in 1998. Archive services also have public workstations, firstly to provide access to their own networked finding and retrieval systems. Internet connection is appropriate at such points as a means to provide remote access to electronic archives and the Webpages of other archive services. Certain types of restriction on access to archive content are already a normal aspect of archival practice. They generally arise from the protection of official and institutional secrecy and of the privacy of individuals (Council of Europe, 1998). However, a strong principle of archival work is concerned with maximising access to documentation of the past in whatever form it occurs, so the provision of network public access points fits into the archive ethos very naturally. 2.3 Cultural institutions It is now a common experience to visit a museum and find some form of multi-media presentation relating to the exhibitions and collections, but many museums also give access to networked cultural resources from outside their own walls. There is also a wealth of access provided by cultural institutions such as medialabs, community arts centres, and small arts-related companies of various kinds. This type of access varies greatly from example to example, but it tends to have in common the fact that it offers more than just the facilities to use content created by others, it provides opportunities to create new cultural or educational content. A few of the large and growing number of such centres are Artec, London; Society for Old and New Media, Amsterdam; Public Netbase, Vienna; Terravista, Lisbon; Mikro, Berlin (examples from an unpublished paper by Marleen Stikker, Society for Old and New Media). Any kind of restriction of access would be wholly alien to the philosophy of such centres, and would seriously inhibit that freedom to innovate with new media which is the centres' raison d'etre. 2.4 Kiosks and cafes Finally, there is access which is public, in the sense that it is available at low cost to anyone who wishes to use it. This is at the same time often private, in the sense that it is provided by private organisations or on private premises. Internet cafes, or cyber cafes, are becoming ubiquitous in big and medium-sized cities. They offer cafe facilities with the possibility of time on networked computers (at a small cost), rather in the way that newspapers and magazines are still offered (free) for the use of customers of cafes in some European countries. There are also public information terminals, or kiosks, in the foyers of public buildings, shopping malls, transport centres, and also in the streets. Most of these offer dedicated access to some network providing local government, consumer, travel, or other similar information. Touchpoint in the UK is an example of such services. However, the Dutch national telecommunications service KPN Telecom now provides street terminals at which Internet ac |
