ABSTRACT
In several Internet situations the requirement is to identify undesirable activity on a site, such as might be carried out by fraud-sters, terrorists, or paedophiles. An example of the latter is a newspaper report of a teenage girl ending up in a dangerous situation having agreed to meet offline an apparently innocent contact made during a chatroom conversation. The contact turned out to be a male predator. Several companies and agencies are now concerned to find ways of identifying potentially dangerous content within the discourse of chatrooms and social networking sites. 1 The dangers have increased following the increased provision of content via mobile phones. Parents at least had the opportunity to monitor online activity when this took place through the home computer, but this opportunity is lost when the contact is made directly to a child’s mobile.
