THE FAIR ARTIFICIAL INTELLIGENCE EDUCATIONAL SYSTEM
University of Malta (MALTA)
About this paper:
Conference name: 13th annual International Conference of Education, Research and Innovation
Dates: 9-10 November, 2020
Location: Online Conference
Abstract:Education is currently facing various challenges. Teachers acknowledge that each student is unique but teaching methods are targeted towards the whole class making it difficult for teachers to cater to the needs of individual students. The Fair Artificial Intelligence System (FAIE) system was designed to provide a personalized learning experience. This is achieved by making use of an Artificial Intelligence (AI) backend which constantly monitors the students and creates exercises which suite the capabilities of those students. It also gradually increases the difficulty of those exercises in such a way as to handhold the student through their learning journey. In actual fact, this is similar to Lev Vygotsky's theory of scaffolding and the related concept of the zone of proximal development. These are teaching methods that can help students learn much more information and much more quickly than they would with traditional instruction. The only difference is that rather than having a human assistant supporting the student, it is the AI that is providing this support. The FAIE browser-based application is aimed at both teachers and students and was specifically designed to be used on their tablets. It generates personalized Mathematics classwork and homework worksheets for primary school students aged between 8 and 10 years. Students can immediately view the results of the completed worksheets and unlock trophies. Teachers may use the app to create their own exercises or utilize the preloaded curricula. The FAIE app was tested in various schools on a total of 280 students. Students were randomly divided into 4 test groups; one control group and the other three groups used the app at school and/or at home. Teachers used the Blueprint Designer and Teacher Interface sections of the app. Both students and teachers answered separate questionnaires regarding their experience using the app. When the pre-test and post-test results were compared, it was noted that low performing students who used the app benefitted the most, with a 23% improvement (on average), while there was no significant difference (no improvement) between the test results of the Control Group. From the questionnaire responses, it resulted that 72% of students preferred working out the worksheets on the FAIE app rather than on paper. The teachers who answered the survey agreed that the app is easy to use and a useful resource. Even though we are in the first stages of development of this system, the initial results are very promising.
Keywords: Education, Artificial Intelligence, Scaffolding, Zonee of Proximal Development, Machine Learning, Vygotsky.