Integration of Artificial Intelligence into the Formative Assessment Process for Students

Abstract

Artificial intelligence (AI) is gradually permeating into higher education, reshaping the way students learn and are assessed. Yet, the concepts of formative assessment as a contemporary model for assessing students are still relatively often neglected. That ́s why we sought to know the reasons for under-effective application of AI in education, and also possibilities for AI application in optimized formative assessment. Methods of induction, deduction, synthesis, mathematical modeling as well as comparative and bibliographical analyses were employed in the study. The study findings identified the good and the bad impact of AI on education and gave recommendations for eliminating negative outcomes imposed in education as a result of AI integration. We categorized a list of popular AI systems in the market by their characteristics, and educational functions. Potential of AI in formative assessment was highlighted and optimization for formative assessment with AI were discussed. Concerning threats, the low level of professionalization of teachers to work with AI, the raising of risks about academic dishonesty and possible unfair student evaluation were mentioned. To minimize AI-negative consequences, we suggested expanding the training of teachers in working with AIs, creating high-quality programs based on AI technology, personalized teacher training in working with AIs at universities, and ensuring access to test programs for detecting AI abuse. As far as potential uses of AI for the improvement of formative assessment, they suggested auto generated testing, trend analysis and recommendation for customized learning/anti-cheating using them. The authors have put forward a mathematical model for evaluating efficiency of the AI tools used in formative assessment that would enable authorities within higher education institutions to adopt AI tools considering the influence of AI on PSD as; diminishing teachers’ labor intensity raised by their quality, academic tasks produced per AI and financial availability of introducing.