Measuring Global Hatred Online
The January/February 2005 issue of Foreign Policy featured a story that I find fascinating because of the issues it raises around how one measures hate on the internet globally. Scholar Stephen P. Cohen and software entrepreneur Jacob Levy have created a bit of code called, "Global Hatred Index (GHI)," designed to gauge levels of hatred online.
From the article, written by Leslie Palti:
"Using algorithms and text-analysis technology, Cohen and Levy’s software automatically searches, scans, and analyzes millions of Internet messages—from online discussion forums, chat rooms, blogs, etc.—to assess levels of hatred toward, say, the United States, Jews, or Muslims. For instance, in one recent index, conducted between June 2003 and May 2004, the software analyzed nearly 2 million messages. The results? An average of 35 percent of the messages expressed some level of anti-Americanism, 32 percent were anti-Semitic, and 20 percent contained anti-Islamic content."
From my view, it was only a matter of time before someone came up with a piece of software that could perform such a task. The emergence of the "Global Hate Index" (GHI) raises all sorts of interesting questions ~ including who will pay for such a program (presumably there's a business model behind it if one of the developers is an entrepreneur). It also raises new possibilities and methodological questions about what can be done in terms of content analysis of the web, in addition to questions about what gets coded as hate speech.
Beyond the issues of who will pay, what the code is, what the methodological implications for studying hate online are, the GHI suggests a techno-cratic approach to combatting hate online. If we count the number of hate messages online, are we closer to stopping them from appearing altogether? Missing from this kind of strategy are the consumers, and the creators, of hate online. What's really needed are new ways of critiquing and thinking critically about hate online. I'm not sure there's an algorithm for that.
From the article, written by Leslie Palti:
"Using algorithms and text-analysis technology, Cohen and Levy’s software automatically searches, scans, and analyzes millions of Internet messages—from online discussion forums, chat rooms, blogs, etc.—to assess levels of hatred toward, say, the United States, Jews, or Muslims. For instance, in one recent index, conducted between June 2003 and May 2004, the software analyzed nearly 2 million messages. The results? An average of 35 percent of the messages expressed some level of anti-Americanism, 32 percent were anti-Semitic, and 20 percent contained anti-Islamic content."
From my view, it was only a matter of time before someone came up with a piece of software that could perform such a task. The emergence of the "Global Hate Index" (GHI) raises all sorts of interesting questions ~ including who will pay for such a program (presumably there's a business model behind it if one of the developers is an entrepreneur). It also raises new possibilities and methodological questions about what can be done in terms of content analysis of the web, in addition to questions about what gets coded as hate speech.
Beyond the issues of who will pay, what the code is, what the methodological implications for studying hate online are, the GHI suggests a techno-cratic approach to combatting hate online. If we count the number of hate messages online, are we closer to stopping them from appearing altogether? Missing from this kind of strategy are the consumers, and the creators, of hate online. What's really needed are new ways of critiquing and thinking critically about hate online. I'm not sure there's an algorithm for that.
posted by Jessie Daniels at
3:58 PM
