Links to: EFF/GILC statement on
"voluntary" ratings; and the EFF
domain naming(ICANN) Archive.

Esther Dyson's comments submitted on September 10, 1999 to the Summit to Discuss Global System for Rating Internet Content:

This response is my personal reaction to the Bertelsmann Foundation Memorandum on "Self-Regulation on the Internet." I applaud the Foundation's attempts to deal with a tough issue, and this response is intended helpfully, not destructively. Like it or not, many forces want to "regulate the Net," whatever that means, and "self" regulation is probably better than regulation by some "non-self" authority. However, it's not clear who the "self" is to be in this case.

Overall, the document leaves me feeling distinctly queasy. So much of it defers details for implementation later. Such and such must be done: Who will do it? Illegal content: Sure, we're all against illegal content, but who decides what is illegal? There are too many questions left open to be answered by some legitimate authority later on.

The basic problem is that the group is attempting to come up with a global solution, topdown. But the nature of the world is that it is a collection of sometimes interacting communities, not a single global administration to be governed top-down. In some spheres there is need for coordination and collaboration, but it does not necessarily need to be governed globally.

Of course, the idea is that the system for Net content regulation would be run by well-meaning, enlightened individuals who know what is best for everyone. But what happened to the notion that people know what is best for themselves and their children? What happened to regulation by citizens themselves of the content they choose for themselves or their children, rather than regulation by a "self" of industry entities beholden to their governments?

The document proposes the creation of a full, broadly integrated set of institutions that can "protect" us all from the problem of illegal content. I fear that we will end up with a worldwide bureaucracy always forced to take the "safe" route, calling for the removal of questionable content. ISPs are properly relieved of responsibility for actions against their customers; let the worldwide content-rating system take the heat. It will take the heat, and dismiss it, because after all it is protecting the public, and a few mistakes here and there are inevitable.

Now let me consider some details:

Illegal content

What is illegal content? Throughout the document, the writers (not named) refer to "illegal content such as child pornography." If there is any content other than child pornography that they think should be illegal, the authors should have the courage to specify what they mean. They make one broader reference: "·.racist and discriminatory web sites, child pornography material exchanged in certain newsgroups and chatrooms and 'how to'-guides on terrorist activity are too disturbing to ignore. Mechanisms have to be developed to deal with illegal content, to protect children online as well as free speech." But that is all.

Later, there is some expectation that "illegality" of content will be determined in the home territory of the publishing website, and will be taken down in accordance with that territory's laws - and presumably by its law-enforcement officials. But in the world of the Internet, with mirror sites, anonymous e-mail and the like, this may not be feasible - fortunately!

The proposed rating system

The proposed rating system, with its three layers, is nicely designed. The idea is to encourage sites to rate themselves, using some common vocabulary, and then to encourage second parties to create rating "templates" with combinations of various metrics and that vocabulary to reflect their values. Finally, a third set of raters should make specific whitelists of acceptable sites - acceptable to children, mostly - beyond the more abstract criteria of the second layer. In theory, that neatly eliminates the value issue from self-rating.

However, First, in the more detailed rating section, the authors propose that the vocabulary be created by an international group of experts of high integrity: "In addition to experts on civil liberties and Internet policy, the board should include social scientists who can advise about what kinds of content are more and less harmful to children." What are these social scientists doing defining a value-free vocabulary? Surely they belong only in the second layer.

Second, a global vocabulary is inherently limiting and too constrained. It's a matter of emphasis, but the value is in the third layer, where people make editorial choices. Otherwise, where's the appreciation of quality, of a sense of humor, respect for the truth? Surely children need to be protected from bland junk as well as from trashy or harmful junk.

Moreover, the focus on protecting children seems excessive. Perhaps it is this focus that makes the idea of almost universal filtering politically palatable. Surely people will have other motivations for filtering, but they might not want to use a filtering system as blunt as this one. Personally, I'd like to see a rating system for truthfulness, for disclosure of advertising relationships, for bias, for political leaning, for assumed audience. (Is the site for techies or for consumers?)

At least there's a provision that unrated sites would not automatically be excluded by most filters.

Child pornography vs. children viewing pornography

The report seems to gloss over the distinction between child pornography, a legal term that connotes the use of children in pornography, which is (almost) universally illegal. This generally involves abuse of actual children, and content on the Net is evidence of the actual abuse of children. This is quite different from the viewing of pornography (on or off the Net) by children, which is almost certainly harmful in excess (like almost anything in excess) but is quite a different matter.

Privacy issues

I also have some concerns over the report's attitude to privacy protection - and implicitly, to anonymity. It is important to catch criminals, but we need to maintain a balance among society's various needs. There is a suggestion that the Internet industry (broadly defined) should "tak[e] all commercially reasonable steps to verify the identity of subscribers, while protecting subscribers' privacy." That does not seem to be necessary: Abusers can be shut off without their identities being known; persistent abusers will eventually become identifiable.

Constructive criticism

So what are alternative, positive approaches? First of all, private groups such as the Bertelsmann Foundation are doing the right thing by getting involved in this debate. They should raise people's awareness of the issues and encourage them to think for themselves - and to pick content for themselves and to offer content-rating services or choices to others. Bertelsmann should encourage private groups and companies to develop and promote rating services, not just for porn or violence, but for quality, advertising disclosure, data-collection-and-use practices, and the like. These services, like many services designed to "solve problems," are a huge business opportunity.

Bertelsmann should also encourage widespread consumer-education campaigns, led not just by foundations and governments, but also by companies (known as "advertising"). Just as consumers look for price, nutritional information, fabric content, care instructions, warranties and other information on products, so should they be encouraged to look for similar meta-information on websites.

In short, let's look at the role that informed, empowered citizens can play in keeping the Net a place they want to live in.

Please send any questions or comments to webmaster@eff.org