BESS vs Image Search Engines

An anticensorware investigation by Seth Finkelstein

Abstract: This report investigates how N2H2's censorware treats several image search engines. N2H2/BESS turns out to blacklist popular general image searching sites as "Pornography". The cached images held by such sites are probably the cause. This is discussed as a deep dilemma of censorware, where general search facilities may be considered as tainted if they contain any cached forbidden material.

Introduction


If good, why do I yield to that suggestion
Whose horrid image doth unfix my hair,
And make my seated heart knock at my ribs,
Against the use of nature? Present fears
Are less than horrible imaginings:

-- Shakespeare [Macbeth, Act 1, scene 3]


N2H2 is a company which makes censorware (aka "filters"). Censorware is software which is designed and optimized for use by an authority to prevent another person from sending or receiving information. N2H2's product is sometimes sold under the name BESS, The Internet Retriever .

This report is concerned with investigating how N2H2's censorware treats several search engines which attempt to find images (not text) related to the desired search term. Note these search engines are not making the hoary claim of image-recognition . They don't pretend to understand the content of the images, but rather work from material such as the name of the image, the surrounding text, or links on the associated page or to the associated page. Thus, the searcher might be presented with wildly varying content based on the search term. See, for example an overview from RLG DigiNews

In order to rapidly return results, and have users quickly sort through the contents, such search engines typically store reduced-size "thumbnail" versions of other website's images on the search engine's own servers (yes, there is a copyright-related issue here, involving "fair-use" , but that issue is beyond this discussion). This is trouble for censorware. Since the content of the images varies widely, and the search engine's images come from the search engine's own server, what is a censorware program to do? Remember, contrary to any overhyped claims of AI , computers have no understanding of content. So a censorware program has a dilemma of either permitting all such images to be viewed, or broad banning.

All information given below can be verified by using N2H2's single-site blacklist checking form . Either type in the URL, or use links referencing the form below. In the following discussion, it will be assumed that anything blacklisted as "Pornography" is forbidden for the user. Some of the particular information below might change after this report is released. But the dilemma described will always remain.

More on Image Search Engines - "pluck it out"


And if thine eye offend thee, pluck it out: it is better for thee to enter into the kingdom of God with one eye, than having two eyes to be cast into hell fire.

-- Mark 9:47


Google

As already discussed in my earlier report BESS vs The Google Search Engine (Cache, Groups, Images) :

BESS quite straightforwardly blacklists all of Google Image Search as: "Pornography" . As some results might contain "adult content", N2H2 then handles that by forbidding everything related to image searching on Google.

Lycos

But, per the above discussion, this blacklisting problem is not restricted to Google's Image Search. Similar issues impel similar treatment of other image search engines. For example, BESS outright blacklists Lycos Multimedia Search as
The Site: http://multimedia.lycos.com is categorized by N2H2 as:
Adults Only
Pornography

It's critical to understand the implications here. Lycos Multimedia Search is by no means a search engine dedicated to pornographic content. It is a general search engine that returns material across the web. But because of the mere potential that it could act as a way of viewing forbidden content (via thumbnail images or similar), it is completely prohibited in the above blacklisting.

Ditto.com

The treatment of Ditto.com is even stranger. BESS blacklists Ditto.com as: Nudity Pornography Swimsuits. This is odd as Ditto.com claims to be quite strict, stating:
"Sites, web pages and images containing pornography, explicit sexual content, and other offensive material are always blocked from search results."
And further :
By adhering to a strict filtering process our web site has received recognition as a safe site from the following web filtering vendors:
Though N2H2 is not among the vendors listed. Perhaps there's no honor among censorware vendors.

[update 1.1, unblacklisted as Pornography, others still in effect, see below]

Altavista

The Altavista search engine features an image search page . Although BESS has this page merely as a Search [Engine] item, doing the search is near-useless under N2H2's bans. It turns out that all of Altavista's reduced-size images are stored on dedicated servers (e.g. "http://thumb-1.image.altavista.com") and BESS blacklists Altavista's image servers as: Pornography

So, when doing an Altavista search, a user can see where the search results would be, if BESS would allow the user to see them. It's a bizarre outcome. Nominally, one is permitted to do the search, but the results are virtually destroyed.

FAST/alltheweb.com

Another full site blacklisting is in effect for http://multimedia.alltheweb.com/ , as BESS blacklists FAST Multimedia Search as: Pornography. But interestingly, their image search facility can be reached as an option from the very general http://www.alltheweb.com/ page . A search for "anderson, pam" proved quite revealing.

[update 1.2, now completely blacklisted, see below]

Conclusion


(C) CERTIFICATION WITH RESPECT TO ADULTS.--A certification under this paragraph is a certification that the library--
(i) is enforcing a policy of Internet safety that includes the operation of a technology protection measure with respect to any of its computers with Internet access that protects against access through such computers to visual depictions that are ...

-- The (misleadingly named) Children's Internet Protection Act


Although N2H2's BESS censorware is examined above, the dilemma is a fundamental problem of censorware. Image search engines are not "pornography". They are general searching tools. However, the mere fact that they have the potential to be used as a way to see forbidden material presents reasons for censorware to prohibit or cripple them.

The standard discussion of censorware tends to be framed in terms of "filtering", that is, removing toxic material. But censorware is in fact about control. It is concerned with making an escape-proof blinder-box. Any facility for escaping this control is a threat to the censorware.

More concretely, when people comment regarding censorware, too many often sneering or snidely frame the debate in terms of why anyone should want or need to read bad, worthless, wasteful, toxic material. Putting aside the civil-libertarian matters just for the sake of discussion, it is questionable how many such people would blithely endorse the extensive censorware prohibitions on general image search engines because of the perceived toxic taint.


Version 1.0, March 5 2002

Update 1.1, March 8 2002 - N2H2 changed the blacklisting of Ditto.com , to remove Pornography (still has both Nudity and Swimsuits though).
But I've saved a screenshot of the original Ditto.com full blacklisting
It's nice to know that someone is reading!

Update 1.2, March 9 2002 - N2H2 now blacklists the entire alltheweb.com search engine as Pornography. Again, this is a general search engine, blacklisted completely as Pornography, because someone might use the image search feature and view prohibited content from the alltheweb.com internal image cache.

Update 1.3: Some time after the above was publicized, N2H2 removed the whole-domain Pornography blacklisting and "merely" blacklisted all image searching as Pornography


Mail comments to: Seth Finkelstein <sethf@sethf.com>

For future information:   subscribe    to   Seth Finkelstein's Infothought list    or read the    Infothought blog

(if you subscribed a few months ago, please resubscribe due to a crash)

See more of Seth Finkelstein 's Censorware Investigations