B.C. Civil Liberties Association http://www.bccla.org 425 - 815 West Hastings Street Vancouver, B.C. Canada V6C 1B4 (604) 687-2919 fax: (604) 687-3045 e-mail: email@example.com FREEDOM OF EXPRESSION ON THE INTERNET May, 1994 Prepared by John Westwood, Dale Beyerstein, and Martin Hahn A. Introduction If the media can be believed, there is a real danger, against which our laws and police are powerless to protect us, lurking behind every personal computer connected by a modem to the telephone line. For example in a recent article originally published in the Ottawa Citizen, luridly titled "Cybercrime: As the information highway grows, so do the terrorists, vandals, paedophiles and other criminals who use it" we read: Cyberspace is a lawless world....Vandals, thieves, terrorists, pedophiles and murderous thugs ride this electronic frontier, knowing law enforcement officers are far, far behind. Police have a shortage of expertise, equipment and some would argue, laws to back them up. Computer users of any age, with a little bit of know-how, have access to hate propaganda, hard-core pornography, stolen credit card numbers, even a massive blueprint to circumvent telephone billing systems. (Reprinted in the Vancouver Sun, April 30, 1994, B5) The picture presented here is frightening. But how well does it represent the reality of computer networking? We shall maintain that this picture grossly exaggerates the problems posed by computer networking. Thus the claim mentioned in the article that we need new laws to handle severe problems with computers is simply not warranted. It is indeed the case, as has been noted by many observers, that present laws restricting speech are difficult to enforce on computer networks. However, the bccla takes the position that these laws should not exist in their present form anyway; so the fact that they are unenforceable on computer networks is more of a blessing than a problem. There is, however, another type of situation where it is arguable that changes are required in the law to provide needed enforcement of court decisions on those involved in computer networking. This case is exemplified by the Ontario ruling in the Karla Homolka case banning publication of evidence. The judge's ruling was enforceable only in Canada, and U.S. media were free to report this evidence. And Americans who participate in discussions of current affairs on discussion groups on computer networks were free to do so. The trouble arose because several of these discussion groups were accessible to Canadians in Canada through the internet. Now, if the reader thinks that the judge's ruling in this case was wrong, then that person will not be greatly offended by the result of this case: that Canadians who were interested got to find out some of the reasons behind a judge's sentencing in this particular case. Nevertheless, the example is disturbing for other cases where the reader does agree that a judge's decision to ban publication is a just one -- say, in cases under the Young Offenders Act. We deal with this case later. But first, some information about how these computer systems work may be helpful. To begin, we should note that we are discussing systems which allow a person to communicate with other people, or extract information from other computers, by connecting her own computer to other computers by the telephone lines. (The device which allows this connection to work is called a modem, and is a feature that must be added to standard home computers.) The systems which coordinate various computers with each other are either (electronic) bulletin boards (BBS), the university (or college) mainframe computer to which those connected to the university in some way get access (sometimes by paying for it) via their home computer, or by FreeNet. These systems serve two important functions in their own right: to allow members to communicate with each other by leaving messages on the system for each other; and to allow for the circulation of programs from one user to another. That is, User A may write a program to do a particular task, and wish to share it with other users. He will upload it to the bulletin board, and any interested user may download it for use on her machine. Bulletin boards come in various types. Many are operated by private individuals on their personal computers in their basements, at no charge to the user. There are over a hundred such systems operating in Vancouver at present. They accept people as users by taking their applications when they call in on their computers, getting the user to choose a password, and usually vetting them by a voice telephone call. Other bulletin boards are run by institutions such as Airspace, an anti-smoking lobby group, for their members or those who share a similar view about smoking. Similar to this are companies that run bulletin boards as a way of allowing employees to effectively communicate with each other, or to work from home. Again, they limit access to a relatively small group, and do not consider themselves open to the public. As well as these, there a number of bulletin boards that charge users a fee. Some of these have gone beyond one city. The largest of these, CompuServe, is worldwide, and is like the Internet, discussed below, in its scope. FreeNets, which now exist in several Canadian cities, exist not only to provide people with the services already mentioned for free, but also to provide terminals for those people who do not own their own computers. These systems also serve another, increasingly important, function. They serve as gateways to the Internet, a system which connects about a million computers worldwide, allowing people to retrieve data from remote computers. Obviously, the procedures for doing this are very complicated, requiring passwords for security, and expert knowledge. Bulletin boards make it possible for relative computer novices to make use of the Internet to extract information. Users of a local bulletin board, as we said, trade messages and files containing programs or pictures with each other. Sometimes they send messages "live" to each other, akin to sending messages back and forth to one another on a teletype. But usually they do so by leaving a message for the other to read at some later time. It is important to distinguish two types of messages. One type, electronic mail (e- mail) is, as the name suggests, a message sent privately from one user to another. The other type of message is not necessarily to one particular person, but if it is, it is still meant to be seen by many others. These messages are posted in what are variously called echoes, conferences, forums, etc., depending on the bulletin board. As the names suggest, these are electronic places where people get together to talk about special interests, hobbies, or current events. There are literally thousands of such areas where one may post a message to start up a new discussion or contribute to one that is already taking place. The Internet has allowed these discussions to become worldwide. One can join a discussion group on a local bulletin board in which messages are carried via the Internet and gathered together from all over the world. It is this latter, trans-national feature of these systems which cause the problem alluded to above in our mention of the judge's ban on dissemination of information about Karla Homolka. Of course the Internet is a new way of transmitting information; and this means new problems with respect to dealing with this information. But does it pose special problems requiring special legislation? Before answering this, we lay out the BCCLA's position on freedom of expression generally. B. The bccla's position on freedom of expression Our position is derived from the writings of the American philosopher Alexander Micklejohn. It could be called "the argument from democracy", and it goes like this: A democracy is a society which is organized in such a way that the citizens, collectively, are the sovereign rulers. The citizens are the ones who, in the end, determine what the laws and policies which regulate our common behaviour should be, and so it is the citizens who determine what the shape and tenor of our society will be. It is essential to our role as sovereign citizens that we have access to the widest possible range of ideas, ideas on any matter whatsoever that touches on our common goals and aspirations. Such ideas would obviously include those in the areas of politics, economics, social welfare, and so on but would also include ideas about sex, race, religion, morality. What is essential is not that each of us have considered all such ideas -- though perhaps that would be a good thing --but rather that there is nothing barring our access to such ideas. In a democracy, the state -- that is, the government -- is the machinery that we use to govern ourselves. We authorize the government to enforce the laws and policies which regulate our common behaviour, laws and policies enacted by our elected representatives. Thus, as well as being sovereigns, we are also subject, in that we are subject to the laws and policies enacted by our representatives and enforced by the state. But there is one area where we cannot allow the government to regulate our behaviour, and that area is our access to ideas. Consider the consequence where the government is allowed to control our access to ideas, To the extent that government is allowed to control our access to ideas, government controls the minds of the citizens. It controls the agenda for which ideas can be discussed in the public forum. In so doing, it has taken over the sovereign role that we as citizens are supposed to have. Anyone who controls the range of information to which we can have access controls the scope and nature of the public debate about how our society should operate. Therefore, so long as we are committed to a democratic form of self-government, we cannot allow the state to prevent our access to ideas. The present point here is not that in a free exchange of ideas, the best ideas will survive exposure to criticism and evaluation -- the cream will rise to the top (although we might wish to defend the value of open discussion for this reason elsewhere, and discussions on the forums of the internet could be seen as examples of this notion at work). The point is, rather, that uncontrolled access to ideas by sovereign citizens is fundamental to, or constitutive of, our democratic way of life. Any withdrawal from that ideal is a withdrawal from democracy. Now it may be that as we survey what passes for the public debate of important ideas, we are not so sure that democracy is a wise course to follow. When we examine some of the ideas that have gained favour amongst citizens, it is sometimes hard to believe that the future of Canada ought to be left in their hands. The same worries will be generated by an examination of the wide variety of messages posted on the various electronic forums, from "alt.anarchy" on the Internet to the White House Forum on CompuServe. The type of person attracted to computer networking is often someone with eccentric opinions, and without any reservations about expressing an opinion on a subject far removed from his or her formal education. Nevertheless, one cannot help but be impressed with the willingness to be educated, the tolerance for outr opinions, and the respect for the rules of argument that are the norm rather than the exception on these forums. On the whole, the discussion on these forums gives one hope for participatory democracy. The B.C. Civil Liberties Association is committed to democracy. Democracy is not a perfect solution to the problems of social organization, and maybe in Canada we're farther away from perfection than we could be. However, democracy is a good deal better than any of the alternatives. And freedom of expression is necessary to make democracy work. C. Weighing freedom of expression against other goods Freedom of expression, and democracy which it serves, are not absolute values. There are eight areas in which the state censorship of ideas ought to be considered: (1)the use of children in the production of pornography; (2) the products of real, non-consensual violence (as opposed to the mere depiction of this) in a sexually explicit context. The reason for allowing censorship in these two areas is that the production of such expressions involves criminal behaviour. The state is justified in banning expressions where it is essential to the content of those expressions that criminal behaviour be recorded and displayed. That is, the justification for censorship is derivative upon another reason; viz. the prevention of other activities which are wrong. Restricting (1) or (2) reduces the incidence of the criminal behaviour in those cases where it is undertaken with the intention of profiting from the sale of the depictions of it. Note that only a small proportion of sex or violence is engaged in for these reasons; so it is not reasonable to maintain that censorship is going to eliminate much of this criminal behaviour. Nevertheless, it will eliminate some of it; and this is a sufficient reason in the eyes of the BCCLA to justify censorship here. There is another rationale for censorship which is direct, in that the information censored is morally or legally objectionable in its own right, and not as a result of some prior objectionable activity. Several types of case involve this type of rationale. (3) the conveying of information that violates an other's privacy. Examples of this would be the posting of stolen credit card numbers, or confi dential medical information obtained in contraven tion of employees' oaths of confidentiality or statutes. Clearly there are grounds for preventing these invasions of privacy no matter the forum in which they occur. Our question, then, will be how to do this with the least disruption to the things we value on the internet. (4) judge's rulings designed to promote a fair trial or protection to a minor. This area has already been mentioned; and will be considered later. There are two other areas in which regulation should not be opposed in principle: (5) Material viewed by children It is appropriate that children should not have as wide an access to ideas as adult citizens. It is not completely wrong-headed to judge that because children's minds are in the process of being formed, and because they may not have had the experience or possess the balanced framework with which to assess certain strong or controversial ideas, their access to such ideas should be limited. This limit on access might take the form of legisla tion restricting the purchase of magazines, or the viewing of movies to those over a certain age. Our present questions are whether there should be, or even can be, analogous restrictions on material carried on computer bulletin boards. We shall return to this topic. But first we look at a type of case which calls for something to be done, but much less needs to be done in these cases than many people advocate: (6) Offensive material encountered without warning It is also appropriate for society to judge that people should not be subjected against their will to ideas which they may find repugnant or offensive. Thus, for example, banning a huge bill board over the Burrard Street Bridge which contained a hate message, or a sexually explicit photograph or drawing, is not unreasonable. Similarly, it may be argued that there ought to be the equivalent of zoning regulations on the internet. Note how this differs from censoring, i.e. preventing access to material generally. When we return to this topic our concern will be to find the best way of determining that there are adequate warnings or placements of such material on the Internet. We are opposed to the censorship of this material solely on the grounds that it is offensive Finally we look at two types of case where censorship has been proposed in recent discussions: (7)speech which could be viewed as incitement. Incitement occurs in circumstances where there is no realistic opportunity for the content of the speech to be cooly considered, its pros and cons weighed, and a rational judgment made on its merits. Rather, the circumstances are such that the speech is likely to directly cause action. Yelling "Fire!" in a crowded theatre is a commonly cited example. Yelling "Lynch him!" to an angry mob holding a brutal serial rapist might be another. What is crucial in such examples are the circumstances in which the speech act occurred, not its contents. The very same content, if delivered in circumstances where time for deliberation was available, would not constitute incitement, and so should not be banned. It is clear that messages read on the computer, whether e-mail or ones publicly broadcast on the Internet will not constitute incitement in this sense. One cannot immediately do anything after reading such a message; and usually the next message in the thread (a continuing discussion) will provide rebuttal. Therefore censorship of the Internet on the grounds of incitement is unjustified and will not be considered further. (8) Material prohibited under the "hate" provisions of the Criminal Code, and the like In E. we shall maintain that this provision does not justify censoring material on the internet. D. When should ideas be regulated? The types of cases mentioned in the previous section show that the proposition that citizens' free access to ideas is fundamen tal to democracy does not entail that the distribution of ideas should not be regulated or classified (as in cases (5) and (6)) or even completely censored in the extreme case. However, such measures must be applied only when they can be demonstrated to be necessary. Demonstrating this, we maintain, requires establish ing two separate points: First of all, we must show that there is some competing good, of greater value, that would require the suppression of freedom of expression in order to survive. We concede that this can sometimes be established in cases of types (1) (4) considered above in C. But second, we must show that the suppression of the expression is the most effective way of preserving that other good. Given the nature of the Internet, we maintain that there are alternatives to complete suppression which are effective enough to protect these goods, and which will not produce as serious a damper on freedom of expression. There fore these alternatives to total censorship should be used instead. These forms of regulation do not constitute unacceptable censor ship, since none of them prevent adult citizens from having access to the ideas or images which are regulated. For much the same reason as (2), the classification, as opposed to regulation, of material destined for public viewing is well within the democratic blueprint. There is much to be said for people being able to know ahead of time whether a film is going to contain offensive or repugnant material, so that they can avoid this material, if they choose to. Similarly, classification on computer bulletin boards could serve the same function. Our question will be "Who ought to be responsible for this classifi cation?" Should it be the government, in the form of laws en forced by police, the owners of the bulletin boards, Internet services and the like, or the individuals, or their parents, that are consuming the services? Help with the answer to this question comes when we consider the second point raised at the beginning of this section. That claim was that in order to justify interference with freedom of expres sion we must show not only what we showed in III above, that there is a compelling reason for restricting the material in a given case, but also that the interference will have the effect of reducing the problem we set out to solve. John Stuart Mill (On Liberty p. 15) sets out two tests that must be satisfied before interference with our liberties could be justified. We paraphrase: (a) Given the mechanisms of control available to the state, would the person or institution to be con trolled behave better because of the sanction than he or she would if not regulated? (b) Would the exercise of this control bring about greater evils than the ones it was designed to prevent? A positive answer to both of these questions, according to Mill, must be "yes", in addition to the demonstration that there is a real evil that needs to be prevented. We shall maintain below that, given the way the internet is organized, the answer to both these questions is no, when we consider specific legislation applying to the internet. That is, the passage of new legislation governing the Internet will bring about evils greater than the ones it is designed to prevent. However, laws currently in place can readily be applied to the internet without being a cure worse than the disease. And, more important, there are non-legislative solutions for some of the problems we have been discussing that are even better answers. E. Applications to electronic bulletin boards or networks The issue at hand is how, if at all, this account of freedom of expression in general applies to electronic bulletin boards or networks. For brevity, we will use the term "bulletin board" to apply to all such electronic media which allow users or members to place messages, files (including those that display images) in the medium, and to access them. We use the term "network" to refer to a connected grouping of such bulletin boards. First, messages on conferences, echoes or forums on various networks are not private communications. Their intention is to be public; that is, they are to be read by anyone who finds the subject heading interesting. The same point holds for files uploaded to a library or file area: the whole purpose of up loading them is that they be shared. They can therefore be regarded as public communications, or expressions in the public forum, and are therefore subject to the same rules as any public expressions. Libel, consumer protection and protection of privacy legislation apply to internet and bulletin board postings The third case we mentioned above, the invasion of privacy involved in posting someone's credit card number or personal information illegally obtained, poses a different sort of prob lem. There is justification here for the state to deter the posting of these types of messages or files, or to ensure their removal once they have been posted. However, the anticipation of these sorts of problems does not justify any sort of prior censorship or control of the bulletin boards. Following the principles enunciated earlier from Mill, what is required is the method that is least onerous that minimizes this problem. What this is here is to leave the regulation in the hands of the people running the bulletin boards, rather than setting up government agencies to scan BBSs for such material. The simple fact is, despite the media attention to this, is that there is very little of these invasions of privacy on BBSs. The cele brated cases of credit card numbers reported have largely been of high profile executives of corporations involved in disputes with environmental organizations and the like -- they are politically motivated, as opposed to being done for some private gain such as profit or revenge. Thus there is not much criminality to police. And, as we mentioned in A., the majority of sysops are hobbyists operating for free, who simply cannot afford to pay fees to pay for such policing, and who cannot afford the risk of time and money that would need to be spent for defence against heavy- handed prosecutions. An important defence of our freedom of expression is defending these BBSs from harassment. The same points hold for those BBSs run by lobbying groups and FreeNet. We shall address some concerns later about how self monitoring of these systems could be effective. We also defer until later a discussion of information which defies a judge's ban. As for the other types of cases cited in various discussions, such as instructions on how to circumvent telephone billing systems, we view these as being blown completely out of propor tion. What is contained in this information is the very same information that is contained in library books, courses in electronics, and other sources that are readily available to anyone. After all, the worriers about this problem could exer cise an immediate calming effect on themselves by simply asking where the hackers got the information they posted in the first place! Thus there is simply no need to invoke any law to handle this problem. The other type of case to have grabbed media attention of late is that of pedophiles establishing contact with young children over computer bulletin boards, and arranging for a meeting in person at some other time. But again, it is important to keep this issue in perspective. The time spent at the computer is not the problem -- contrary to the hype issued by the computer gurus about cybersex. What is of real concern is the meeting later. But no one would seriously blame the telephone or the mail where these were the source of the original contact with people parents think undesirable. Parents understand the need to monitor their children's magazine subscriptions or telephone contacts, and to exercise control over their children's behaviour while away from the home or parents. The computer is no different in these respects from the telephone or mail. Granted, bulletin boards offering interactive games are more attractive to many children than are magazines. And many parents are ignorant of the workings of, and afraid of, the computers they bought for their children. Nevertheless, it remains the parents' responsibility to monitor their children's use of these instruments. No law could possi bly be effective in supplanting this role, and where parents are fulfilling the role, laws are unnecessary. Granted, parents cannot lean over their children's shoulders every moment they are at the computer; but software is presently available to restrict access to certain places where certain subjects are discussed. Unfortunately, Parliament and the B.C. Legislature have not been persuaded by the bccla account of freedom of expression. Messages on a network could be subject to Criminal Code sanctions, and to sanctions under the Human Rights Act. For example: It is likely a Criminal Code offense to possess and/or distribute some of the material which could appear in networks. * s. 318 prohibits advocating genocide; * s. 319 prohibits promoting in a public place hatred against an identifiable group; * s. 163 prohibits the distribution, and possession for the purpose of distribution, of obscene material; * s. 163.1 prohibits the production, possession or distri bution of child pornography (which includes counselling or advocating sex with someone under 18). A conviction under the Criminal Code could lead to jail. As well, B.C.'s Human Rights Act prohibits the public display of any representation which is likely to expose a person or class of persons to hatred or contempt. A proceeding under the Act could lead to an order to financially compensate the complainant. Finally, it may well be that any of the originator of a message or image, the moderator of a forum, or the operator of the network (the sysop) could be sued in civil court for defamation or for invasion of privacy. It is possible that there is something about the electronic transmission of images on a network which would preclude it from prosecution under the Criminal Code or Human Rights Act. If so, lawyers could have field day with this material. But even if there are legal loopholes, it is likely that Parliament and the Legislature would soon act to close them. State censorship of ideas has been confirmed twice recently by the Supreme Court of Canada, and it is likely to be the law for some time to come. Nevertheless, it is our position that these sanctions have no place in a democratic society; and we therefore oppose their extension into electronic media. Our position is that self regulation by the sysops and conference moderators themselves, to prevent the sorts of invasions of privacy or other genuine concerns has worked up until now, and there is no justi fication for legislation that would negatively affect freedom of expression. F. Self Regulation We now turn to the question of how the self regulation of expres sion on networks would work in practice. Both sysops and desig nated members of a network who moderate the conferences or forums have the power either to delete material or to classify it -- that is, place it in files where special passwords are needed in order to access the material. In most cases there are practical difficulties involved in requiring the sysop to be responsible for all the material on his bulletin board first of all because of the large quantity of material which some networks contain, and second, because of the fact that a great deal of the informa tion gathering on a typical bulletin board is automated. A typical bulletin board runs unattended a good part of the time, and messages are posted, and even removed after they have been on the system for a reset time without the sysop doing or monitoring anything. Also, many of the messages on a typical bulletin board are generated elsewhere, and passed along the network to all the bulletin boards that take that conference or forum. On a typical network such as FIDO Net, the computer running one bulletin board will automatically dial another computer to receive a very large volume of files containing individual messages, files and the like, retain some of that information for the bulletin board, but merely pass on some to the next computer that will call it at a preset time. Information travels across North America by one computer calling another in this way (programmed in so as to eliminate or minimise long distance calls). All the decisions involved in this transfer of information could have been made months ago; and the sysop of a particular system would have no need to scan any of the information that ended up on his or her bulletin board, let alone the large volume that "passed through" that computer. This would suggest that the main burden for regulating these messages for such matters as libel or general offensiveness would have to fall on the moderators of the conferences, since these are the only people whose job it is to actually read the mes sages. However, these moderators are usually volunteers, and cannot be expected to be aware of all the laws and statutes that might apply in all the various jurisdictions where the conference is picked up. For that matter, they are often not even aware of all the localities where the conference is available. Also, because they are volunteers, moderators will often resign, and conferences will run unmoderated for varying lengths of time. As well, there are several conferences on the Internet that pride themselves on being unmoderated. This problem, however, is by no means as serious as it looks. Our position is that the number of messages that should be weeded out is very small. Therefore the need for constant vigilance on the part of the moderators is not great. And we ought not to underestimate the moral sense and acuity of the other readers of messages on the system. Where something truly offensive slips by, such as the posting of someone's credit card number, other readers will immediately send a spate of messages to anyone who looks like they may have some authority to do something about it. A similar set of considerations apply to judges' bans of informa tion. The problem here is that the information that works its way onto a BBS in an area under jurisdiction of the court is generated in a place not under the court's jurisdiction. This problem is not difficult to solve, however. The situation here is conceptually the same as a newspaper picking up a news item from a wire service such as American Press. In this case, there would be no difficulty prosecuting the Vancouver Sun for publish ing the item, despite its coming from New York. The same point, seemingly, would apply to BBSs. This solution requires no new legislation specific to electronic media. The practical difficul ties, however, were noted above. Sysops or forum moderators would be required to pay attention to the messages they carry, rather than relying on the automated system. However, these difficulties are not insuperable. The computer provides sophisti cated ways of searching messages for key words such as "Teale", "Homolka" etc., and messages about banned items are likely to show up only in certain discussion groups having to do with laws or current affairs. Those BBS operators who have no interest in these areas do not carry them, and have no problem. Those who do as a service to their users will have to undertake a bit more caution when a judge's order is made. However, we should empha size that in a well run society the number of orders that a sysop would have to take cognizance of would be extremely small. This is not an unfair burden to place on those who wish to provide this service. For those who run afoul of the law in this matter there are already adequate provisions for punishment. But, putting aside practical difficulties, should material -- especially sexually explicit material and hate messages -- be deleted or regulated by private people? Would this be preferable to censorship by the government, with all the safeguards demo cratic governments build into their policies and procedures, and the enforcement of them? Let us first look at the deletion of material. There are at least two types of material that must be distinguished. The first is electronic mail, the second are the messages and files that are intended for the public places on the network. With respect to the first, privacy is the paramount concern of the users of these systems; and therefore most users do not wish to see their messages censored or monitored. However, we note that different bulletin boards offer their e-mail services on different terms; and we think that these terms should be left up to the consumer and sysop. When there is payment for this service, there is an enforceable contract. We think that civil law serves as an adequate remedy for problems here. However, the general principle here is Neville Chamberlain's, "Gentlemen do not read each others' mail"; therefore we think that the minimum of interfer ence in this area the better. Public areas such as file areas or message areas are slightly different. There is no reason why sysops or moderators should not delete material which, for whatever reason, they wish to exclude from access through the network. This situation is analogous to a bookstore owner or gallery owner who makes a personal decision not to carry or display a particular work. They are not "censor ing" the work -- that is, they are not using the power of the state to prohibit the work from being distributed -- they are simply refusing to participate in its distribution. The author or artist is perfectly free to try to sell the work through other outlets. An identical argument applies to sysops. They should not feel themselves bound on principled grounds to allow messages on their network which they find, for whatever reason, offensive or dangerous. The person who wishes to transmit the message is perfectly free to try to put it on another network, or to express it through other means of communication. However, there is a counterargument which deserves mention. Art galleries and bookstores are about the only way that artists and authors can get their work to the public. Some artists and authors are controversial, in that their work challenges and crosses accepted boundaries of taste and sensibility. It is, I think, not wrong to say that they are the lifeblood of the artistic community, for art when it is good challenges accepted ways of looking at ourselves and our world. Can we conclude that art gallery owners or bookstore owners, who profit from the talents of artists and authors, and who represent the only venue for the display of new works, are under a moral obligation to take risks in showing or carrying avant-garde works, even if their paying customers might be offended and stay away in droves, and even if they risk criminal prosecution? We think that such an argument bears some weight. But can it be applied to sysops? We don't think so, at least not to those running small bulletin boards from their own personal computer. One of the beauties of the internet system as it has presently involved is its plurality. When someone is denied access to a particular bulletin board or forum, this will not be the only venue for such messages or images -- there are ample opportunities for displaying such messages in other places. Therefore, this argument cannot be easily applied to sysops. The fee-for-service bulletin boards such as CompuServe are slightly different. They offer a public service, and in fact will usually be charging money to the person whose message they are denying access. Nevertheless, first of all, these services specify very clearly what services they are -- or are not -- offering; and none of the ones presently operating offer their clients unlimited rights to post messages. If they did, that would be another story. Secondly, we might also think of an analogy with letters to the editor in a newspaper. It is not a violation of anyone's rights in the typical case off an editor's decision to refuse to publish a letter. We think that a sysop's, or a moderator's decision to remove a message because of reasons of relevance (wrong discussion group), good taste, or worries about libel are defensible. Even in the extreme case of a moderator removing a message he disagrees with, deplorable as this would be, does not seem to us to be grounds for legislation to monitor these decisions. The case is even clearer for bulletin boards run by particular groups for particular purposes. The BBS for the Atari (Computer) Users' Group should not be required to carry a message about the evils of smoking, and Airspace should not be required to carry a message on how to oil the door on a floppy drive. This is not to say that some sysops cannot choose to provide a very wide latitude in the messages they will allow the network to carry, even a completely unrestricted arena for ideas. They may view themselves as providing a public good in doing so, but they should feel themselves obligated to allow extreme messages, at least not in the same way that art gallery owners may have such an obligation. What about the regulation and classification, as opposed to outright prohibition, of messages on a network? Does this run afoul of any principled protection of freedom of speech? The answer seems to be no. As we mentioned earlier, the bccla accepts a range of means for regulating access to expressions. These include restrictions on the public display of controversial material, restricting access by children, and alerting potential viewers ahead of time of the nature of the content of a work. These are already in place in the "real world" for films, videos, magazines and books. There is no reason why the justification for such restrictions on access to material shouldn't apply to networks. The issues are identical. And, in fact methods of regulation have already grown up on the internet. Moderators of conferences and forums regularly remove messages from one message area when they are clearly more relevant on another (e.g. when a discussion of homeopathic remedies for dogs surfaces on an alternative health forum for people). However, here it is almost universally the case that the message is forwarded, e.g. to the pet health forum, rather than destroyed. Moderators also on occasion will restrict access to someone who is rude or offensive to another person on the forum. It is only a small step to extend this practice to other sorts of offensive messages. Self regulation would thus seem sufficient to handle the problems raised by those calling for government censorship. As for keeping children from accessing certain message areas with inappropriate material, the technology provides two solutions that can be implemented by sysops that are already used widely. One solution is for the sysop to provide different users with different types of access to the system when a person is accepted for membership. Thus a minor who is accepted onto a particular bulletin board may simply be denied access to certain file or message areas where this material is stored. The defect with this approach is that it cannot distinguish between different family members who access the system with the same password. A better solution is to secure certain areas of the system with a separate password that must be entered before access to that area is given. Critics of bulletin boards point out that there is no direct contact between the sysop and the user; and so the sysop cannot monitor whether a person logged on to the system is the original adult who was granted access or his 9-year-old son who has found out the password and is logging on when the adult is not around. While this is true, the same problem exists with the distribution of magazines that are restricted to those over 18. Even if the shop owner is especially careful to sell these magazines only those over 18, how can she be assured that they will not later fall into the hands of a minor? We do not think that this possibility justifies an outright ban on these maga zines. In fact, the situation is probably better with respect to computer bulletin boards. Whereas the most immature young person can open up a magazine and look at the pictures, it requires a certain level of maturity to operate a computer and log on to a bulletin board. Granted, the kind of maturity involved in being a computer wizard is different from that required to properly appreciate pornography; but at this point we reiterate our comments from V. that it is the parents' responsibility to control the computers in their private homes, just as it is theirs not to leave inappropriate reading material in the wrong places. And this brings us to a third solution, which can be implemented by parents. Software is already available which blocks access to certain discussion groups by name, or which screens messages with certain content. Thus parents do not have to breathe down their children's necks every instant they are at the computer in order to have control over what their children are reading. Now granted, this software is not any more sophisti cated than Bowdler in making selections; but it certainly rules out a great deal of what parents usually want. The gap between these programs and parents' judgment about a specific message ought to be bridged by discussion about general principles that is central to parental guidance. As a means of directing children to appropriate material, and for warning adults to avoid material they would find offensive, the internet provides a kind of classification for messages. Discus sion groups with titles such as "alt.conspiracy.kennedy" make it very clear what is being discussed there. (If the cryptic title doesn't give it away, almost all forums have a statement which members read when they first sign on telling them, sometimes in very explicit detail, what kinds of messages are welcome there.) And messages within these discussion groups are not displayed to the browser on the system. What is displayed is a short header, such as "Oswald CIA payments". This header may be used by someone responding to the original message, and the person responding to that, and so on. (A continuing discussion using the same header is called a thread.) Thus casual browsers can be warned of potentially offensive messages; or, once having identi fied messages in a particular thread that are offensive, readers will know not to look at other messages with that heading. With files the system provides an even better classification: files are given descriptions usually up to 5 lines long describ ing their contents. Files that display an image that some may find offensive are not immediately displayed on screen; viewers will search by these descriptions and only then go through a fairly complex and time consuming procedure to view them. In fact, because time generally costs the user money, the usual procedure is to download these files to the home computer and view them later. Doing this requires special software, and an understanding of how to use it. A combination of extraordinary incompetence with improbable bad luck would be necessary for someone viewing an image by accident. The only problem is a misdescribed image. And, given that the knowledge and skill necessary to intentionally download and view such files is reasonably great, and the correlation of these with maturity is positive, though not by any means perfect, we submit that anyone who can get the image up on the screen is not an obvious candi date for paternalism. And where paternalism is appropriate, it ought to be parents practising it rather than the state. After all, the computer the child is using is in the home. The other sorts of files that raise concern are text files that can be downloaded. These consist of threads of previous ex changes of messages on topics someone though would be of inter est, or long files -- sometimes rejected books, or parts of books that never made it into the published version. In principle, these should be treated no differently than messages: that is, classification and warning, as opposed to legislation and polic ing are what is required.. G. Conclusion Our position is that the public cry for policing of computer bulletin boards and networks is misplaced. None of the eight categories mentioned in C. where censorship might possibly be justified turn out to require special legislation to handle. (1) and (2), violent or sexual images produced using chil dren or non-consenting adults, is already covered under legisla tion dealing with the criminal actions themselves. (5) and (6), material offensive to some adults or inappropriate for children, is best dealt with by the self censorship mechanisms outlined in F.. (7), incitement, is, as we argue in C., not a problem on the internet. (8), material prohibited by criminal code and human rights legislation, we argue in E., should not be prohibited in the first place. This leaves two categories that are problematic. (3) material invading others' privacy, such as posting someone's credit card number, is a serious, though infrequent problem. Given its infrequency, current legislation can be applied to those bulletin boards that make no effort to remove these items once they are brought to their attention, and to those individuals who post the material. Critics of our position will maintain that it is very difficult for authorities to enforce the current laws on the internet. This is, of course, true. But the same point applies to any new laws anyone might propose. And if these new laws erode the goodwill of bulletin board users and owners to police them selves, they will only make matters worse. Last, (4) the enforcement of court publication bans, is again a serious problem. But we maintain that the same mechanisms for enforcing these against wire services apply to the Internet. And, as we have already argued, the legitimate publication bans are few enough that they can be enforced against the very few who defy them. In conclusion, then, those who advocate extending state powers of censorship to this new area should first do a better job of demonstrating why it is needed where it presently exists. Our position is that this job simply has not been done.