How can video-games deliver educational content in an intelligent fashion?

This paper describes a research proposal which explores the feasibility of utilizing cognitive and motivational scaffolding techniques from tutoring systems, in the development of video-games. The motivation behind this project is that we could build a video-game to tackle difficult school topics for children in year 5. We believe that the use of tutoring systems’ techniques for cognitive and motivational modelling could enhance the “intelligence” of the video-games so that these adapt to particular children offering personalized experiences in the context of a video-game that could also deliver educational content. Introduction and background The employment of video games in educational settings has been thought to be motivating for children (Malone 1980). With this idea in mind some video-games such as “Zoombinies” (Hancock and Osterweil 1996) or “Prime Climb” (Conati and Zhou 2002) have successfully exploited the video games’ appeal to deliver educational experiences to children. On the other side of the spectrum, the use of Intelligent Tutoring Systems (ITS) has tried to emulate an educational situation aiming at delivering appropriate feedback to improve the children’s performance in the use of the software and arguably, in the children’s knowledge of the topic. The nature of the ITS’s feedback depends on the approach taken for its construction. The provision of help has been an important element in the definition of such systems. Luckin and du Boulay (1999), for example, have operationalised Vygotsky’s philosophy of teaching and learning by incorporating an implicit “more able partner” that suggest, but not directs, the type of help the ITS “believes” the child needs. The system’s beliefs are based on artificial intelligence techniques that dynamically model the learner’s cognitive state regarding helpseeking. In the last decade, a great deal of research has produced ITS’s that have considered the learners’ motivational and emotional state together with the cognitive state (Hudlicka and Fellows 1996; De Vicente 2002; Chaffar and Frasson 2004). Del Soldato and du Boulay (1995) for example, developed a set of motivational rules taken from theories of motivation and video games, to produce motivating feedback that could encourage the learner to put more effort or take more challenging activities. Another example of the use of emotions in agents has been presented in Lester, Towns et al. (2000). The provision of such feedback is underpinned by models of the learners’ motivational states that, similarly to cognitive models, are based on artificial intelligence techniques. The efficiency of ITS that are both ognitively and emotionally intelligent has been studied producing some interesting effects (Lester, Converse et al. 1997; Rebolledo-Mendez 2006).