Today, the OpenNet Initiative has released a paper, authored by Helmi Noman and myself, enumerating the widespread use of American- and Canadian-built filtering technologies in the Middle East and North Africa. The paper, entitled “West Censoring East: The Use of Western Technologies by Middle East Censors 2010-2011“, looks closely at Websense, McAfee’s SmartFilter, and Netsweeper in Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, Sudan, Tunisia, the United Arab Emirates, and Yemen, and particularly at how websites–including my own, the OpenNet Initiative’s, and Amira Al Hussaini’s blog–have been mis-categorized by these technologies, resulting in what is essentially censorship.
While I suggest you read the paper (or this excellent Wall Street Journal article reporting on it), I’d like to discuss briefly why my blog was categorized as pornography by Websense. Frankly, I find it utterly fascinating: About a year ago, Helmi Noman–my co-author–discovered that this very blog was blocked in Yemen. Upon further investigation, Helmi realized that the reason for the blockage was not political content or anything of the sort, but that my blog had been categorized–by Websense–as pornography.
After Websense barred Yemen from future software updates, I thought the problem had been solved until Luke Allnutt–who works at RFE/RL, which uses Websense in its offices–tweeted that he couldn’t get to my blog.
I quickly wrote to Websense, and received a fairly rapid reply, telling me that my blog had been reclassified as a personal site. Great–I then pushed back a bit, asking how my blog had been categorized as a pornographic site in the first place. My assumption was that their automated system was based on keywords, and that my blogging about Helmi Noman’s paper (“Sex, Social Mores, and Keyword Filtering: Microsoft Bing in the ‘Arabian Countries‘”) had caused it; after all, it caused “Arab sex” to be the #1 search term for my blog.
Turns out, that wasn’t the case at all. In fact, what happened was significantly more chilling. Here’s the text of an email sent to me by Patricia Hogan, Senior Public Relations Specialist for Websense:
Regarding your questions about blog classification, the problem seems to come from the comments, not the posts. Indeed, you appear to be the victim of comment spam (which often contains pornographic links or links to malware).
Look at the comments after this post: http://jilliancyork.com/2008/09/11/blog-strike-for-mohammed-erraji/. The last comment has pornographic links and the one preceding it has links to pharmacy spam, which often leads to malware. This is just one post that we looked at. You may have more.
Comment spam has been hounding bloggers (and more recently Facebook users), so Websense developed tools to help keep blogs and readers safe from spam like this. We offer free plug-ins for many blog platforms to help prevent this type of comment abuse (go to http://defensio.com/downloads for more information). We don’t want you to be victimized again from unscrupulous posts, and our plug-in allows you to control what content you wish to appear on your site.
I hope this helps. Please let me know if you have any more questions.
Sr. Public Relations Specialist
What Hogan is saying is that anyone can manipulate Websense software by spamming a blog’s comments section with porn outlinks. Let me say that again: Websense can be manipulated by anyone wishing to censor anyone else, just by adding a few links to porn in the comments section.
SmartFilter appears to have similar problems. A few months ago, blogger Sabina England reported that her blog was blocked in the UAE, which uses the software. While she may have a similar issue with “porn spam,” our suspicion at the time was that SmartFilter was detecting keywords, and had blocked England’s blog based on the use of the words “cunt,” “sexy,” and “whores” in a poem she had written.
I find this utterly chilling; now, I will say that Yemen has stopped using Websense and we’re not aware of any other countries–at least in the Middle East and North Africa–that use the software. Nevertheless, plenty of schools, libraries, and workplaces use Websense and other tools, and while their blocking of pornography may be justified, the mis-categorization of URLs by these technologies means that there are chilling effects, even to blocking porn.