Most major news outlets have included some sort of comment service for their website, so that readers can place comments on any articles that are posted. However, moderating the comments posted by every man and his dog can often involve armies of moderators, especially for a site like Huffington Post.
In the case of the Huffington Post, an army of 30 human moderators in combination with software are employed to moderate the nearly 30 million comments the site receives every month. The Huffington Post finds this a major issue to the extent that it had made its first ever acquisition by buying Adaptive Semantics.
Adaptive Semantics is the developer of an application called JuLiA which uses ‘supervised machine learning’ to reduce the workload of human moderators to something more manageable without removing comments which normally add to the debate around a particular news piece. Other companies currently using technology from Adaptive Semantics include CNN, Newsweek and Disqus. By purchasing Adaptive Semantics the Huffington Post will exclude their competitors from using this technology – the news site will honour existing contracts but does not plan on renewing them.
Beyond the moderation of comments, the semantic analysis technology can also move the best comments up the list, and the Huffington Post thinks that there are also further applications for the technology. One idea is to award ‘badges’ to users who regularly comment on certain topics, e.g. Afghanistan, so that other users can identify and follow the opinions of ‘experts’.
However, semantic analysis is not the only way towards successful moderating. Websites such as Slashdot and engadget use creative user moderation techniques in an attempt to achieve the same goal. For the users at least, the effort pays off as quality debate is encouraged while abusive comments are kept off the radar.