'Governance of controversial Internet content in the European Union.'
5 Protecting Minors from Adult Content
The issue which has stimulated the most concrete proposals from the Commission is that of access of children to adult material which may be harmful to their development. The type of content here primarily involves sexual material and violence. All of the key policy statements support the use of filtering and rating, and in particular the PICS system. The Communication, for instance, envisages a Council recommendation 'setting out a clear political message encouraging the use of filtering software such as PICS, and for one or more European rating systems' (Commission, 1996a). Likewise, the Council's Resolution on the issue calls on member states to 'encourage the provision to users of filtering mechanisms and the setting up of rating systems for instance the PICS (Platform for Internet Content Selection) standard' (Council of Ministers, 1997). The other key documents all mirror this stance. Filtering software has been the main technological response to the problem of children accessing 'adult content'. The emphasis the Union has placed on this area reflects, in my view, the sense among policy makers that this is one of the more viable policy instruments on offer. It also reflects disquiet at the amount of pornography available on a medium which represents such an economic and cultural resource.
Yaman Akdeniz, in his study of Internet pornography regulation (1997) outlines the different forms of availability. These range from the pictures, short animated movies, sound files and stories found on the World Wide Web ('WWW') to sex related discussions on the Internet Relay Chat ('IRC') channels where users in small groups or in private channels exchange messages and files (Akdeniz, 1997). There are also Usenet discussion groups devoted to sex, and Wayne Myers (1996) outlines how the Internet can make it possible to see live sex acts and arrange sexual activities from computer screens. While the proportion of web pages containing explicit content may be very low, in Joel Weinberg's estimation 'an astonishingly small fraction' constituting 'less than a tenth of one per cent' (1997: 1), these pages tend to be well flagged in search engines. The point is, a child only has to search under the first taboo word that comes to mind, in order to access a plethora of adult content. I have noticed in my own researches that fairly innocuous search terms yield copious hits pointing to explicit material (e.g. 'toys', 'love' and ironically, 'PICS'). And while Akdeniz (1997: 3) claims that 'most of the WWW sites with pornographic content require proof of age and payment by credit card to access their materials' my own researches quickly revealed a huge range of free samples and many completely free sites, easily accessible by any minor, and graphically representing every mainstream and 'paraphiliac' category I had known to exist, and some more besides (though thankfully no child pornography).
The technological response of creating filter software for installation on the computers of end users has been an inevitable attempt to recreate 'gatekeeping' mechanisms that have existed in more traditional media. The global nature of the Internet makes any 'watershed' viewing time impossible. Similarly, the is no gatekeeper analogous to a retailer obliged by law to refuse to sell 'top shelf' materials to minors. Internet filtering software has received wide acclaim as a means of blocking the access of children to content that may be harmful to their development. Plantiffs in ACLU v. Reno (US District Court for the Eastern District of Pennsylvania, 1996) relied heavily on the existence and capabilities of blocking software in arguing that the CDA was unconstitutional, while President Clinton has pledged to 'vigorously support' the development and widespread availability of filtering software, mirroring the European Union position. Free speech activists see such software as providing the answer to the dilemma of indecency regulation, making it possible 'to reconcile free expression of ideas and appropriate protection for kids' (Daniel Weitzner quoted in Weinberg, 1997: 1).
The European Commission's 'Communication' (1996a) usefully distinguishes three main models of filtering software. These are the 'blacklisting' model, where access to listed sites is blocked; 'white listing', where access is only possible to listed sites; and 'neutral labelling', where sites are labelled or rated, but it is up to the user to decide how to use the label or rating.
The first generation of filter packages generally follow the blacklisting model, and I will argue, suffer from serious shortcomings as a result. Programs such as Cybersitter, Surfwatch and Netnanny rely on blocking access to sites on the blacklist and sites containing certain blacklisted 'strings', that is, keywords. The blacklist in some cases is downloaded and updated online, and is difficult to access. In others the list is accessible in the software's configuration settings. These packages are stand alone systems, which may be purchased and downloaded online. Besides the effort and expense involved for parents, the fact that these programs are not integrated into the web browser means that they can be easily deleted by the precocious minor - unless protected by another program designed to deny access to selected files.
The author has installed one of these packages (Net Nanny) on an experimental basis, and has seen some of these drawbacks exemplified. I accessed just one of Net Nanny's banned sites and discovered it to be a completely innocuous page about skiing. Apparently, some programs block entire directories of web pages because they contain a single 'adult' file (Berlin and Kantor, 1996). This means that large numbers of innocuous Web pages are blocked merely because they are located near some other page with adult content, which probably account for the unfortunate fate of the skiing page. Some programs block entire domains, including all of the sites hosted by particular Internet service providers (Meeks & McCullagh, 1996).
Another snag arose after my first installation, when I found it impossible to access EU documents of relevance to this thesis. It transpired that 'pornography' itself was a banned word. The only other banned word was 'anarchism'. Apparently vigilant parents must also take into account the private political agendas of software manufacturers when choosing a filter. While with Net Nanny the blacklist of banned words and banned sites is easily accessed, manufacturers of other programs keep theirs highly secret. What arises as a result often amounts to censorship, with no accountability. Cyber Patrol blocked animal-rights web pages because of images of animal abuse, including syphillis-infected monkeys, classing these as 'gross depiction' (Weinberg, 1997: 8; Meeks and McCullagh, 1996).The problem was aggravated by the fact that Cyber Patrol, following the entire-directory approach described above, blocked all of the hundred or so animal welfare, animal rights and vegetarian pages hosted at the Animal Rights Resource Site. Anything remotely connected with homosexuality is liable to be blacklisted by Cyber Patrol, Cyber Sitter and Net Nanny (Weinberg, 1997: 9). Banned directories include ones with 'vital information from the Centers for Disease Control and Prevention, the AIDS Book Review Journal, and AIDS Treatment News' (Meeks and McCullagh, 1996). Weinberg outlines the often 'surprising and alarming' nature of blacklisting. He instances Cyber Patrol's blocking of Usenet newsgroups including alt.feminism, soc.feminism, clari.news.women, soc.support.pregnancy.loss, and alt.support.fat-acceptance. The same program blocks the Web site of the League for Programming Freedom (a group opposing software patents), and the Electronic Frontier Foundation's censorship archive. Cyber Sitter blocks the National Organization of Women web site, and sites which contain the words 'Sinn Fein' (Ó Marcaigh, 1997). Disturbingly, it blocks the site of Peacefire, cyber-rights group run by teenagers, purely on the basis that it contains criticism of CYBER Sitter. The makers of Cybersitter have even threatened to block every site hosted by the service provider which hosts the Peacefire site (Weinberg, 1997; Ó Marcaigh, 1997). In the words of Meeks and McCullagh (1996), who procured leaked blacklists, '[t]he smut-censors say they're going after porn, but they quietly restrict political speech.'
Other problems arise from the sometimes ridiculous antics of string-recognition software. America Online's software, for example, in seeking to screen four-letter words embedded in text, refused to let users register from the British town of 'Scunthorpe' (Ó Marcaigh, 1997; Weinberg, 1997). The same company, enforcing a rule forbidding certain words in personal member profiles, barred subscribers from identifying themselves as 'breast' cancer survivors (Weinberg, 1997: 15).
The shortcomings in these first generation filtering programs are summed up well by Yaman Akdeniz (1997: 23):
..sometimes this kind of software goes too far away and limits access to or censors inconvenient web sites, or filters potentially educational materials regarding AIDS and drug abuse prevention. Furthermore, the companies creating this software provide no appeal system to the banned content providers, thereby subverting the self-regulating exchange of information that has been a hallmark of the Internet community.
Contrasting with first generation filtering software, PICS is aimed at providing a standard infrastructure for 'neutral labelling' and filtering Internet content. It separates the two functions of rating of sites and filtering of sites (and, significantly, of individual pages). Praised for allowing a high degree of flexibility and security, PICS is acclaimed by the European Commission as 'undoubtedly the most comprehensive and innovative solution yet to tackle Internet contents issues.'(Commission, 1996a). PICS tags Internet sites with, in the words of the Commission, 'value neutral' labels. The Commission gives an excellent, albeit idealised, account of PICS in operation:
These labels can support different types of information: ratings (for instance, evaluating language, nudity, sexual content, violence), or pointers (identifying contents according to their relevance or interest for various constituencies of users). To be viewed, the site must (1) carry a PICS label, (2) be within the parameters set by parents on the home computer. Ratings can be established by content providers themselves (such as entertainment companies operating family-oriented web sites) or by third parties (such as religious groups or parents' associations). Each family decides which ratings systems it wishes to use and then, using the parameters, what is acceptable and what is not. (Commission, 1996a).
While PICS can theoretically support different rating systems, any PICS compliant rating system must be rule based. The rating system most frequently associated with PICS, almost to the point of being synonymous is the Recreational Software Advisory Council's RSACi (i for Internet) system (Akdeniz, 1997:23). Table 5.1 summarises the RSACi system, and it is easy to deduce from it that the correspondence of 'levels' with numerical values is the technical key to the PICS. RSACi parameters were originally devised for rating computer games. The PICS protocol, according to the European Parliament 'would appear to be the most comprehensive and ground-breaking means of dealing with the problems of content on the Internet'. Is PICS capable of living up to the high expectations of its European Union supporters? Below, I will look at some of its inherent limitations, and the obstacles which must be overcome if it is to become a global standard.
| Level | Violence | Nudity | Sex | Language |
| 0 | No aggressive violence; No natural or accidental violence. |
No nudity. | Romance, no sexual activity portrayed. | Inoffensive slang; no profanity |
| 1 | Creatures injured or killed; damage to realistic objects |
Revealing attire | Passionate kissing | Mild expletives or mild terms for body functions. |
| 2 | Humans or creatures injured or killed. Rewards injuring non -threatening creatures |
Partial nudity | Clothed sexual touching | Expletives; non-sexual anatomical references |
| 3 | Humans injured or killed |
Frontal nudity | Non-explicit sexual touching | Strong, vulgar language; obscene gestures; use of epithets. |
| 4 | Wanton and gratuitous violence. |
Provocative display of frontal nudity |
Explicit sexual activity. |
Extreme hate speech or crude language. Explicit sexual references |
Table 5.1[footnote 2] The Levels and Categories of the RSACi rating system.
Jonathan Weinberg has provided the most important scholarly assessment of filtering and rating systems to date (Weinberg, 1997). While viewing such systems as a 'huge step forward' he contends that:
..it is worth trying to locate the technology's limitations and drawbacks. Blocking software is a huge step forward in solving the dilemma of sexually explicit speech on the Net, but it does come at a cost. People whose image of the Net is mediated through blocking software will miss out on worthwhile speech -- through deliberate exclusion, through inaccuracies in labelling inherent to the filtering process, and through the restriction of unrated sites. (Weinberg, 1997:1)
Weinberg surveys the different types of blocking systems, outlining the distinction between approaches based on 'rules' and those based on 'standards'. PICS may incorporate only rule-based systems such as RSACi. Designers of a PICS compliant system must devise simple, hard-edged rules, with the results 'turning mechanically on a limited number of facts' (Weinberg, 1997:5). According to Weinberg:
Legal thought teaches that rules and standards each have disadvantages. A problem with standards is that they are less constraining; relatively speaking, a standards-based system will lack consistency and predictability... Rules become increasingly necessary as the universe of law-appliers becomes larger, less able to rely on shared culture and values as a guide to applying standards in a relatively consistent and coherent way. (1997:12).
Thus, the W3C and the European Union, in seeking to promote a global standard among heterogeneous cultures is seeking to overcome this diversity through the rule-based PICS. Weinberg affirms this rationale, arguing that with the 'large universe of evaluators' necessary to map the Internet, a ratings system relying too heavily on the standards approach would be prone to great dangers of arbitrariness and inconsistency (Weinberg, 1997:6).
Rules-based systems suffer their own shortcomings. In a rule-based system it is impossible to include determinations as to whether nudity is 'artistic', 'erotic', or 'pornographic'. If we consider the categories relating to sexuality in Table 5.1, we see that there is no attempt to distinguish educational materials from other depictions, so that users can allow the former but not the latter (Akdeniz, 1997; Weinberg, 1997). Such classification requires more a more standards based judgement on the part of the evaluator. According to Weinberg rules based systems 'direct law-appliers to treat complex and multifaceted reality according to an oversimplified schematic' (1997:6). In order to simplify a complex inquiry, factors which a sensitive decision maker would otherwise take into account are 'screened off'. At best, a rule-based filtering system can miss nuances; at worst, it can generate absurd results, as in the cases of banned strings described above. As Weinberg explains:
There is no way, consistent with rulishness, that it can seek to distinguish the serious or artistic from the titillating. It achieves rule-boundedness, and ease of administration, at the expense of nuance; it achieves consistent labelling, but in categories that do not correspond to the ones many people want. (1997:7)
The challenge facing ratings system designers is to devise rules-based systems which encompass the complexity of content as well as possible, minimising these difficulties. According to Weinberg the product of any such effort will necessarily be flawed. Other researchers are stronger in their criticism. Jonathan Wallace asks how he is to rate 'An Auschwitz Alphabet', a deeply chilling work of reportage on the Holocaust. The work contains descriptions of violence done to sexual organs of concentration camp inmates. According to Wallace, a self-rating system would force him to choose between the unsatisfactory alternatives of labelling the work as suitable for all ages, on the one hand, or lumping it together 'with the Hot Nude Women page' on the other. Legal researcher, Yaman Akdeniz raises a similar query concerning his 'Cyber-Rights & Cyber-Liberties' web site, which deals with the regulation of child pornography on the Internet.
How will this site be labelled under the PICS/RSACi initiatives? Hopefully, it is an informative and educational site but sometimes there may be strong language such as in the court cases reported in the newspapers. Some rating authorities may judge the site as an offensive even though it has a public purpose. This is the same for various newsgroups or web sites dealing with serious issues such as sexual abuse and AIDS. (Akdeniz, 1997)
One also thinks of sites depicting human rights abuses. According to Akdeniz, there is no room for dissent in the system because ratings will be done by private bodies rather than by government, leaving an accountability deficit. Such concerns force Akdeniz to conclude that 'further research into this problem may be needed before implementing it as a standard for labelling the Internet content.' (Akdeniz, 1997: 24)
Problems are also encountered in choosing which categories with which to label content. According to Electronic Frontiers Australia (1997), 'the definitions used in determining the four [RSACi] categories were 'clearly chosen with computer games in mind and lack the flexibility required for a wider range of materials. It is ludicrous that such a system should be applied to novels, online libraries, art galleries, and other such resources.' Practicalities demand a limited amount of categories. To keep things manageable, RSACi ignores much content that some other ratings systems class as potentially unsuitable, including speech relating to 'drug use, alcohol, tobacco, gambling, scatology, computer hacking and software piracy, devil worship, religious cults, militant or extremist groups, (and) weapon making...'(Weinberg, 1997). In Weinberg's view, if the Internet is to be rated in a comprehensive manner, fairness and consistency can only be achieved if the ratings system uses simple, hard-edged categories relying on a few, easily ascertainable characteristics of each site. Any such categories will be inherently value laden, despite the claim of neutral labelling.
Compared to the regulation and governance of illegal content, PICS poses very different issues of implementation. No questions concerning liability or intermediaries arise; indeed few legal issues arise. The crux of PICS is the need for universality of rating. As Weinberg explains:
Blocking software can work perfectly only if all sites are rated. Otherwise, the software must either exclude all unrated sites, barring innocuous speech, or allow unrated sites, letting in speech that the user would prefer to exclude. (1997:20)
.
The most obvious obstacle to be overcome is the sheer scale of content. One solution is to encourage self-rating where evaluators follow a rule bound questionnaire following the categories of the rating system, be it RSACi or whatever. There is the inevitable problem of misrepresentation of sites, which could possibly be dealt with by a citizen hotline reporting mechanism, similar to current hotlines for illegal content. But a far greater problem is that of providing an incentive to rate. According to Weinberg the incentives are highly uneven:
Mass-market commercial providers seeking to maximize their audience reach will participate in any significant self-rating system, so as not to be shut out of homes in which parents have configured their browsers to reject all unrated sites. Many noncommercial site owners, though, may not participate. They may be indifferent to their under-18 visitors... For the owner of large archives containing many documents, supplying a rating for each page may be a time-consuming pain in the neck.
Then there are those I have instanced above who may be philosophically opposed to the rating system. Rating may also involve financial costs (Weinberg, 1997:22). It seems likely that any incentive would be likely to increase as the amount of rated pages reaches a certain critical mass, the point at which parents are happy to leave their browser set at a default setting which excludes unrated pages. But reaching such a point would necessitate mass compliance in the first place. Industry backing has meant that the market leading web browsers now come with PICS/RSACi included. At present, even though most newly installed browsers, my own included, contain the rating system, very few sites are actually rated. The result is that if unrated sites are excluded there is little material to view, even for minors. Allowing access to unrated sites leaves one faced with most of the offending content the system is designed to deal with.
2. Abstracted by the author from 'PICS-Version 1.0' in the <rsaci.rat> file which is installed automatically into the Windows operating system with recent versions of Netscape Navigator and Internet Explorer which incorporate the PICS protocol.