A pedagogical agent which incorporates a text mining tool to promote collaborative writing

This paper discusses how mediation strategies can be used to design a pedagogical agent integrated to a text mining tool to promote collaborative work in academic reading and writing activities. The agent uses the Artificial Intelligence Markup Language (AIML) to communicate with students, while its text mining features enables it to keep track of what is being produced collaboratively. Mediation strategies based on sociocultural theory are also employed in order to endow the pedagogical agent with the capacity to intervene in students’ discussions according to precise goals. A specific architecture is proposed in the paper as a form to integrate these different mechanisms in the pedagogical agent. The proposed solution explores the roles of interaction, mediation and assistance as key elements for structuring the pedagogical agent's participation as a peer guide in collaborative text writing.