Three General Approaches for Implementing User Empowerment

The above criteria can be used to evaluate the range of approaches to filtering and blocking technologies. There have a been a spectrum of such approaches offered, with more being developed every day. For convenience, we consider three broad categories of approaches characterizing the evolving area of user empowerment:

• No filtering -- Work toward an unfiltered net: this approach assumes that the evils of filtering and harmful uses of the technology outweigh any possible positive effects. In the filter-free future there may or may not be laws which censor content online.

• Ubiquitous, uniform self-labeling: Move toward a universal, ubiquitous content rating system: full and effective user empowerment will only come with uniform content labels that users around the world can filter based on their own values. Assuming that such labeling was created, labeling could be legally-compelled, left purely voluntary for content creators, or be compelled by pressure from browsers which may block all unlabelled content and search engines which may refuse to search un-rated sites.

• Multiple, Third-Party Filtering: Support a diversity of third-party labeling and filtering systems as a means to provide the diversity of families online with the ability to block, filter based on filters, ratings, and labels assigned by other individuals or organizations that the family chooses to trust.

Along with any of these options will also exist a variety of Internet resources which point users to online content of particular value for children, or others. These positive resources will certainly be critical for the development of the net as a useful resource. Unlike blocking, filtering and label tools, these positive resources do not pose significant risks to the free flow of information from any perspective. The fact that they are not discussed further is not a comment on their utility or importance, merely the fact that they are uncontroversial.

In the analysis that follows, these three broad approaches are evaluated in context of the criteria articulated above. While no one approach is endorsed, we hope that the questions raised herein will provide a basis for a rational discussion of the various approaches to addressing concerns about children's access to inappropriate material online.

I. Option One: The Unfiltered Net

Following the U.S. Supreme Court CDA decision, critics of Internet blocking and filtering technologies stepped up their effort to smoke out the evils of blocking and filtering technologies. Taking these concerns seriously, we must consider that the alternative to having blocking and filtering technologies as part of the net is the absence of these tools.

A. No Filtering: An Effective and Trusted Solution?

By definition, the unfiltered net leaves users, including parents, with less control over the content that comes into their homes than they have today. Advocates of less filtering, or a total absence of filtering, clearly believe that families ought to rely on non-technical means of guiding their children's Internet usage. There is general agreement that healthy, constructive use the of the Internet by children begins with parental involvement and responsibility.

QUESTION: Will an unfiltered net leave parents who have chosen to use some form of filtering feeling less secure and trusting?

B. Sustainability of No Filtering?

Filtering and blocking tools are based on the fundamental building blocks of the Internet infrastructure and do not, in fact, depend on new technology specifications such as PICS. User empowerment tools -- beginning with America Online's Parental Controls and products like SurfWatch and CyberPatrol -- are able to block online content simply by using basic Internet addressing features -- domain names, web addresses known as URLs, numeric IP addresses which are assigned to every computer on the Internet, and newsgroup names. Specifications like PICS are important tools for making filtering easier to deploy, but blocking and filtering tools have been successfully deployed without PICS.

QUESTION: Is the Internet likely to remain unfiltered or will the Internet marketplace continue to provide user empowerment tools in response to user demand?

C. Diversity of Voices and Editorial Choices: Will decreasing filtering options enhance or reduce the diversity of voices online?

Creation and dissemination of blocking, filtering, and other content evaluation tools constitutes important expression of editorial and critical judgments which increase the information available online. Moreover, any steps to decrease access to the tools necessary for large and small, commercial and non-commercial speakers to develop and distribute their own filtering judgments might seriously disadvantage the editorial expression of these groups, leaving only large, commercial, well-financed groups and individuals in the filtering arena.

QUESTION: Does shunning filtering on the net, increase or decrease the diversity of voices online?

D. Inviting Censorship: Will a lack of user control invite or discourage censorship?

The United States Supreme Court decision overturning the Communications Decency Act was, without doubt, a landmark decision for freedom of expression on the Internet, inasmuch as it declared that speech online is entitled to the highest degree of First Amendment protection. The decision to overturn the statute turned on two critical factual findings: first, that the burden on speakers to limit access to minors is so great that it amounts to a total ban on speech, and second, that less restrictive alternatives such as user empowerment technologies are available to parents, so the government need not step in to protect children. From all sides of the political spectrum, policy makers agree that if parents have options that enable them to shield their own from children from material they judge inappropriate based on their own family values, then the government need not step in to make the decision for the parents. All plaintiffs in the CDA challenge presented evidence and made legal arguments stressing the availability of these technologies.

QUESTION: Will decreasing the degree of user and parental control online increase or decrease the pressure for governments to censor the Internet?

Contrary to the concern that a decline in filtering will increase the call for censorship, critics of filtering worry that the availability of filtering tools will actually encourage government censorship. With the tools at hand, the reasoning goes, governments will be emboldened to act. The recent history of Internet censorship suggests, however, that the cause and effect is actually the reverse. Countries like Singapore and China with a political tradition of censorship and repression began Internet censorship programs long before PICS and other filtering options were available. Even more liberal countries such as Australia, that are calling for use of the PICS-based RSACi system, have a long tradition of censoring film and TV. Germany, which has sought to prosecute Internet Service Providers for providing access to illegal political materials (racist and xenophobic statements), rejected proposals to employ user empowerment tools.

QUESTION: Are countries that seek to censor the net able to do so without the help of blocking, filtering and labeling tools?

QUESTION: Has the availability of user empowerment tools, now and in the future, offered countries that option of enabling parents to protect children from what they believe to be inappropriate, without sweeping restrictions on the freedom of expression?

II. Option Two: Ubiquitous, Universal Labeling

At the other end of the policy spectrum from the unfiltered net is the completely labeled net. Several self-labeling system -- most notably RSACi and SafeSurf -- propose to have all content on the Internet labeled by its creator according to a uniform, objective rating vocabulary. With all content thus labeled, users would be able to block access to that content which fails to meet their individual criteria.

A. Effectiveness and Trustworthiness of Self-rating

1. Effectiveness of objective content ratings

The effectiveness of any objective, self-rating system turns on the feasibility of creating a single, uniform rating vocabulary that is sophisticated enough to express adequate information about Internet content to enable families around the world to make filtering decisions, while being easy enough to use and understand that Internet content creators, including individual users, can easily label their content on a real-time basis. In addition to being effective, the system must be trusted. PICS technology provides the basis to protect against fraudulent alteration of labels by unauthorized users, but this does not vouch for the accuracy of the labels.

QUESTION: Will parents filtering based on labels applied by the creator of the content trust that the label is accurate?

The problems of developing an effective, objective uniform self-rating vocabulary have shown themselves most directly in the difficulty RSAC recently experienced applying its ratings to online news sources. The RSAC system allows content providers to rate their content with numeric scores representing various levels of sex, violence, nudity or harsh language. This classification systems is in keeping with its goal of helping parents avoid online content that contains, in the parents judgment, unacceptable levels of such material. Applying these ratings may be relatively straightforward for clearly erotic content such as pictorials from the Playboy website, or for obviously inoffensive material such as a geography site. However, what about the MSNBC or Time Magazine web site? On any given day these sites may contain reports on grisly airplane accidents or murders, discussion of sexual abuse, or reports which contain, for some reason, harsh language.

QUESTION: How should news be rated, if at all? What about political or issue advocacy sites?

QUESTION: Should a news story on a earthquake describing death and dismemberment receive the same rating as the Jean-Claude Van Dam fan club web site?

QUESTION: Who should decide what a bona fide news site is?

As of this writing. most of the major news organizations announced that they would not self-rate their sites and RSAC has put it's "news" label plan on hold indefinitely.

B. Will a uniform rating vocabulary limit diversity of speech online?

Keyed to Western tastes and values, the RSAC system provides tags for sex, nudity, violence and harsh language. But what about issues such as blasphemy, racism, political extremism, excessively democratic discourse. All of these are categories of speech which some individuals or governments might wish to avoid by blocking and filtering.

QUESTION: Can a single vocabulary be developed which captures the concerns of all communities and cultures that participate in the net, now and in the future? Could it at the same time be made simple enough to be used by the millions of content providers (speakers) online?

C. Sustainability: Are there sufficient natural incentives for self-rating without resorting to technical or legal coercion?

In the two years since RSAC was launched, roughly 43,000 sites have rated, or 2-4% of all the total number of sites estimated to be on the net today. For self-rating to work, nearly all sites must rate themselves.

QUESTION: What is the incentive for web publishers, large and small, to rate? How will self rating systems bridge the gap from 2-4% to 100% coverage?

QUESTION: Who will take on the task of auditing labels for the millions of web pages that exist today? Who will pay for this effort?

Some have suggested that critical points in the Internet infrastructure could be altered in order to force sites to rate themselves of become nearly invisible to users.

QUESTION: Should web browsers be configured to block access to all unrated sites?

QUESTION: Should search engines refuse to search sites that did not have RSAC ratings?

QUESTION: Will it be necessary to distort the Internet infrastructure in order to create "incentives" for self-rating?

QUESTION: What kind of burden, if any, does this place on individual publishers?

D. Self-rating Invites Censorship: Lack of incentive to rate suggests government mandates will be required

If the abovementioned infrastructure distortions are not achieved, or are not successful at forcing 100% of Internet content to be rated, it might become necessary to make self-rating a legal requirement. In the United States, such a law would likely run afoul of First Amendment protections against "compelled speech." First Amendment doctrine holds that the government forcing someone to speak -- in this case to label his or her content -- is every bit as much a violation of the First Amendment as a government action which prohibits speech [2].

QUESTION: Would such an effort pass constitutional muster?

QUESTION: In the event some legal pressure is necessary, would US legislative action alone be sufficient?

QUESTION: Would it be desirable or feasible to negotiate an international requirement to rate according to a single ratings code?

QUESTION: Does this create an infrastructure for government censorship that would not otherwise exist?

III. Option Three: Multiple, Third-Party Rating Services

A third general approach can be found in the development of multiple, independent third-party rating services. In the last few years, over fifteen independent third-party filtering products and service have come to market. More commercial products come to market every month. The platform for third-party ratings afforded by PICS may enable the creation of an even greater diversity of third-party rating services, but this process is slow, and without truly encouraging results as of yet.

A. Developing effective tools and building trust among users

Third-party blocking and filtering services for sale as stand-alone software, or packaged as part of commercial online services. Most of the known products on the market work by creating lists of sites that are deemed inappropriate for access by children [3].The producers of these products are continually updating these site lists so that even newly launched sites will be blocked if they meet the filtering criteria. None of the products claim to block 100% of what should be blocked, but do claim to cover the vast majority of sites. Moreover, because there is an active, competitive market for these tools, and because computer and consumer magazines have begun to test the relative performance of various products, market pressures are forcing competing products to do a better and better job of finding sites which should be blocked.

QUESTION: Can third party rating services keep up with the rapidly growing volume of content on the Internet?