the internet has become integral part of our lives and it offers new ways to connect to each other. While this has many benefits, it also has a darker side. Over recent years there has been a consistent stream of news articles about children being put at risk due to unsafe online communications. There has been a global drive to educate parents and children around the dangers and provide them with guidelines for safe practice. Yet still the reports that another child has gone missing after going to meet someone they had only met online persist. In this paper we are presenting our work in designing and developing a prototype system that will provide a safer environment for children online. In order to combat such a complex problem, it is necessary to integrate various artificial intelligent technologies into an autonomous system. The system was implemented using Java Agent Development Environment (JADE), Protégé Ontology development and reasoning techniques, and natural language processing (NLP). We have devised a testing system to measure the effectiveness of the system in detecting potential risk to the child. Initial results of experiments are encouraging and demonstrate the system’s capability in preventing sensitive information, such as name, address and telephone number being transmitted as part of an on-going conversation. The system will be capable of identifying levels of threat in the context of a longer conversation and recognise the key factors, intention, location and time which contribute to meeting arrangements and proactively generating a warning.