Discovery & Invention The British Association Annual Festival of Science Newcastle 2.45pm Tuesaday 12 September 1995 _________________________________________________________________ PROBLEMS IN THE GLOBAL VILLAGE by Prof. Harold Thimbleby Email: harold@mdx.ac.uk World Wide Web: http://www.cs.mdx.ac.uk/harold Middlesex University London Abstract Some people see the Internet and the World Wide Web as an important step towards democracy, education and peace, and of benefit to everyone from children to entire nations. They see Utopia in the electronic 'global village.' The reality is rather different from the vision. The Internet brings pornography and computer viruses; it tells you how to take drugs and make bombs. These things aren't necessarily what you want in Utopia. We'll explain computer viruses and other forms of computer damage. Although we'll say less about porn per se - for obvious reasons - the technical problems of detecting it are the same, even to the extent of different people having different standards of what is acceptable. The talk will show why the dangers are serious and why there can be no effective safeguards. The Internet won't change on its own. Computer viruses Computer viruses are an electronic form of vandalism. They are a form of booby trap and cause destruction without the direct presence of the vandal. They can affect anyone who uses computers. There is evidence that most companies seriously attacked by viruses are out of business within a year, because they lose track of all their computer records: who owes them what and what their stock is. Amongst virus writers there is a dubious argument that viruses and computer hacking improve computer systems by exposing weaknesses. Part one of this talk explains what viruses are and how they work. We then show that viruses cannot be detected, though particular ones (such as Michaelangelo) may be recognised quite easily. Cryptography (where viruses modify and conceal themselves) is now routinely used and is making virus detection harder. Computer viruses will be around for as long as we want computers to be able to do anything. Calculators don't get viruses; but if you want a general purpose computer to run spreadsheets, word processors and games, it can also run viruses and Trojan Horses. Pornography Pornography is an increasing problem for anyone who uses the Internet, and a worry for parents and schools. It should be said immediately that the pornography on the Internet comes from many countries (including ones with different cultures and standards, such as Japan) and is very easily available. Much of it makes "adult shops" in London's Soho look very tame indeed. Material includes high quality graphics, instructions, stories, sounds, movies, shop catalogues, ... - for both conventional sexual interests as well as all variations. Moreover, the Internet supports bulletin boards and interactive chat sessions: these facilities are heavily used. Some people are very liberal about pornography; they often think pornography is limited to pictures of humans posing in titillating ways. There is certainly that on the Internet; exchanging and publishing photographs is commonplace. However I have found text, film and sound material that I find extremely disturbing, for example involving instructions for killing minors. Overtly sexual pornography is not the only problem. The Internet has information on every activity that Amnesty International and, I believe all sensible humans, would wish to eliminate from the planet. What I say here about pornography applies to all other forms of information. Paedophile and other groups use the Internet to organise rings and meetings, for sexual and other encounters. Is it really a problem? Many people who deny this, I believe, have not got full access to the Internet, or haven't tried hard enough to find anything, or haven't got a helpful teenager!, or possibly hold extreme liberal views. Approximately 10% of shops using the Internet sell 'erotica' (figure from First Virtual). Approximately 10% of bulletin boards I can access directly from the University are pornographic (I used stratified sampling, sampling 50 random BBS out of 50000). Of an analysis of searches made via a Web search engine by people all around the world, 47% of the 11000 most often repeated searches were pornographic (like everything else, this is a subjective estimate because, for example, I counted searches for "hardcore" but I did not count searches for "gay" or "lesbian"). There were 17 searches for a certain abbreviation which (as of writing) match nothing, but on other search engines find plenty of sites (the first site on this topic I looked at told me I was the 44963rd person to look at it); such results suggest that users are well aware of what Internet abbreviations to search for. Like the case with viruses there is a strong element of rationalising self-justification, particularly in the US where the constitution permits free speech. There is also an argument that participants should know how to do potentially dangerous activities but safely. A number of bulletin boards are supposedly in aid of victims, but also serve to make new contacts and to share techniques of, for instance, child entrapment. By using graphic images as an easily understood example, part two of the talk shows how easily pornography can be concealed in several ways. However, the discussion applies to any form of information, graphic, textual, animation or whatever. We conclude that there is no reliable way - technical or otherwise - to detect or intercept pornography. Below we give a brief review of partial answers. Technical mechanisms * Self censorship Sites either register themselves as 'adult' or their adult material includes standard keywords indicating it as restricted access. User's machines recognise the restrictions and restrict actual access. This approach is similar to the proposed V-chip for US television censorship. The method has obvious advantages: it is simple and easy to understand. It would limit access to some material. Arguments against it are, in the main, two-fold: it enables parents/guardians to be over-zealous, and therefore perhaps reduce revenue for legitimate uses of the medium (an argument more relevant for TV); secondly, anyone who willingly registers themselves is unlikely to be a problem anyway. * Indirect censorship Users run software (possibly with hardware help) that restricts access to selected Internet sites. Here we don't rely on self censorship, but we have to take on trust that the list of excluded sites is both sufficient and acceptable. This method is used by SurfWatch. The method, as currently implemented, is easy to subvert by local hacking which would certainly not be beyond teenage hackers. Note that it need not be your child who is the hacker - there need only be one person in the playground able to supply the necessary skills. * Blanket censorship There are available simple programs that stop viewing all graphics. Though of some use for text bulletin boards, these programs defeat the value of the World Wide Web for any purposes. Moreover, pictures themselves are probably the least worrying aspects of the Internet's porn. * Rapid response Systems like SurfWatch can be updated automatically. It is therefore possible to subscribe to sites listing exclusions. As soon as an official watchdog detects undesirable material at some site, it can be listed amongst the exculsions. Note that the watchdog may make finer classifications, such as violence, sex, racism, and users of the system may have personal preferences. * Monitors and access logs Systems can summarise what sites have been accessed. This may be useful for parents reviewing children's Internet use, for example. Note that many sites use aliases and code words, if not accessed via anonymous servers, so that the names of sites may not be much help! Some systems (e.g., the Internet Filter) automatically email other people (e.g., the parents) when a specific site is explored. Ironically, doing an internet search for 'surfwatch' or 'internet filter' will pick up pornographic sites complaining about these programs as well as sites describing their actual use or how to obtain them. Of the sites mentioning SurfWatch, half approved of it, half disliked it - half of those intend to publish the sites SurfWatch classifies to make them more easily available. About a tenth use its classification of them as proof that their site is bad enough to be classified by it. * Limited access Many Internet provides, such as CompuServe and America Online, provide simple restricted access to the Internet. * Cryptography Secret code systems (such as PGP, Pretty Good Privacy) are already used on the Internet to protect privacy. As pornographers use these encryption methods for their own purposes (e.g., to evade detection), increasingly their material will anyway become inaccessible to people who don't have the right keys. This may make the problem self-limiting. Authentication (people know who they are, and people cannot impersonate others) is closely related to cryptography. At present, it is easy for a 12 year old to pretend to be 40 or vice versa. Some pages claim to be made by children so young (including their photograhs) the information must have been assembled by older people. Authentication technology (plus national IDs!) makes impersonation harder. Non technical mechanisms * Better material The Internet has very little interesting material; it needs more. When it has more, it is just possible that pornography etc will slip into its statistically appropriate place, one aspect of humanity, but not the most prominent on the Internet. I've made a list of interesting places in London with their own Web sites. It isn't very long. Why isn't there more good material available? There are several 'electronic libary' projects that aim to make books and information more easily available - see Project Gutenberg for example. The British Library has a useful presence; but where is everyone else? If the British Library's books were all on the Web, it would quite rightly contain some pornography. (This is a common pro-pornography line of argument.) But it would also contain a lot of other stuff, and the pornographic material might be more in perspective. There is a huge amount of very valuable information, for instance on AIDS. (In some cases, the valid warnings about infection become covers for explaining activities that arguably do more to increase the problems.) * Clear thinking Many believe the Internet brings automatic benefits, and naturally leads to democracy, peace and general good will. There is little evidence of this, though there one must not forget the 1992 Russian revolution (see an IRC log), Tienanmen Square and a few other notable exceptions. Particularly in the US, freedom of speech is idolised, and indeed it is very easy to ridicule censorship, particularly with 'thin end of the wedge' arguments. Very few people are aware of the proper statistics. A newsworthy incident, say of an Internet organised child abduction, can be made to appear to be a world wide problem. Though child abductions in cars are far more frequent, who'd argue against cars on those grounds? Note, anyway, that statistics from the Internet may under-estimate potential problems: one child in a school playground might access an undesirable site and then distribute information around the school using floppy disc. Thus statistics might give an impression of low activity, whereas there need only be a very few selective people who pre-select material before sharing it with others who perhaps do not even have Internet access. There is evidence that teenagers have attached pornography to computer viruses. The viruses then spread the material automatically, without the original perpetrator being directly involved. There is a lot of news coverage of Internet pornography. A lot of this is driven by voyeurism, so possiblyly over-estimating the significance of the issues. * Education Though the emphasis on the protection of children is narrow (and is often used to displace further argument), it is certainly the case that children have to grow up and live as best they can in the world in which they find themselves. The fact that there are nasty things someplaces isn't reason in itself to avoid those places; many people live in London, for instance. People do need to learn, sometime, to be wise. Many people argue that parental responsibility is the core issue. "You had children, now be responsible for them." This view seriously under-estimates the difficulty of being responsible when children often have much better technical knowledge and access to material on the Internet. Teenagers are often unsupervised, and they are ones likely to search out sexually interesting material. A report in the British Medical Journal shows that boys given sex education at school were less likely to take part in under-age sex, and girls were no more likely than otherwise. More research like this is needed. Conclusion The Internet has been called a global electronic village. If so, most of it is a heavily-used red light district. Too few people are providing any other interesting or useful services on the Internet. "I have seen the future and it needs work" I shall very briefly describe my own attempts at providing interesting World Wide Web material. Further reading (World Wide Web references are indicated throughout this article; they can be clicked on to obtain the information. They are probably underlined or otherwise easy to see on your browser.) "Sex online: what parents should know," similar stories in both Newsweek and Time, July 3, 1995. Marketing Pornography on the Information Superhighway: A Survey of 917,410 Images, Descriptions, Short Stories, and Animations Downloaded 8.5 Million Times by Consumers in Over 2000 Cities in Forty Countries, Provinces, and Territories, by Marty Rimm, Carnegie Mellon University. Note. "Reader Advisory: Due to the nature of the studied material, Minors and Those Who Are Going To Be Easily Offended are hereby warned of strong language and sexual content. Leave now if dealing with such things is going to be a problem for you." A critical review of it by Brian Reid. Another review by Donna L. Hoffman & Thomas P. Novak and a reponse to its criticisms. The future does not compute, S. L. Talbott, O'Reilly & Associates, 1995. See also the review by Harold Thimbleby, to appear in Times Higher Education Supplement. "An organisational solution to piracy and viruses," by Harold Thimbleby, Journal of Systems and Software, 25(2), pp207-215, 1994. ... many popular Internet magazines, such as .net. _________________________________________________________________ [INLINE] Harold Thimbleby is Professor of Computing Research, Middlesex University, London. Phone: +44 181 362 6061 Email: harold@mdx.ac.uk Web pages: http://www.cs.mdx.ac.uk/harold _________________________________________________________________ Computing Science's home pages & Other talks This file is available on the World Wide Web at http://www.cs.mdx.ac.uk/research/village.html _________________________________________________________________