MODELLING AN HOLISTIC ARTIFICIAL INTELLIGENT EDUCATION MODEL FOR OPTIMAL LEARNER ENGAGEMENT AND INCLUSION
While there exists an extensive body of research literature in the area of Intelligent Tutoring Systems (ITS), little advancement has been made in extending existing ITS to facilitate learners with special needs. Depending on the level of disability, need or impairment, students with sensory and/or learning-based disabilities can be excluded from interacting with ITS. The emergence of AI gives an opportunity to provide accessible solutions for students. Any solution needs to ensure that the user does not feel isolated or vulnerable in their differences through impairment. In the last number of years many AI-based solutions have emerged in relation dyslexia, autism and hearing impairment etc. Currently no ITS exists to allow students of different impairments to work singularly or collaborate in learning environments. The inclusion of AI for accessibility could also aid in the reduction of extensive academic emotion (EAE) (Mehigan & Pitt 2019). Emotions can significantly influence student learning and, as a result, achievement (Villavicencio 2013). The inclusion of enjoyable and interesting activities for example, can lead to increased student engagement (Frenzel et al 2007). This engagement is potentially extended to those with special needs where they feel included and equal to peers during collaborative tasks. The incorporation of AI-based accessible features within ITS can help achieve this. Learner effort can also be influenced through personalisation, contemplation of cognitive variables etc. and “the development of customizable and adaptable applications tailored to them [students with special needs] provides many benefits as it helps mould the learning process to different cognitive, sensorial or mobility impairments” (Fernández-López, 2013). Thus, the relationship between academic emotions, accessibility, cognitive variables, content and learning outcome should be a combined consideration within the learning context to achieve optimal and holistic Artificially Intelligent Education Models (AIEds). This could lead to the development of improved mediated models that further meet student needs in overcoming negative emotions and exclusion in the learning process. This paper outlines a brief history of the convergence of the AI and accessibility. The paper will focus on the potential to incorporate such accessibility solutions into an ITS. An existing ITS model called MAPLE (Mobile Adaptive Personalised Learning Environment) is assessed in light of its strengths and weaknesses for achieving optimal learner engagement through the inclusion of AI-based accessibility features. MAPLE is designed for use with any eLearning or mLearning platform to facilitate the intelligent detection of user learning-styles based on two dimensions of the Felder-Silverman Learning Style Model. Based on user interaction, automatic provision of adapted content display to suit the learning style needs of the user is facilitated. MAPLE is compared to a descriptive and simplified educational technology driven by AIEd technology (Luckin 2016). This technology comprises core models which aims to support the social, emotional, and meta-cognitive aspects of learning. However no provision is considered within either model to incorporate accessible features. Appraisal of MAPLE to an AIEd driven model could inform the extension of existing ITS to achieve optimal mediated AIEd to meet the holistic needs of the learner including those with special needs.