Innovative Educational Models for Human Centered AI

Innovative Educational Models for Human Centered AI Vladimir Pavlović1, Nataša Milosavljević1, Olivera J. Bogunović2, Bogdan Novaković3 1University of Belgrade, Faculty of Agriculture, Serbia, vlaver@agrif.bg.ac.rs, natasaglisovic@gmail.com 2 Harvard University, Harvard Medical School, USA, obogunovic@mclean.harvard.edu 3 University of Milano-Bicocca, Department of Human Sciences for Education "Riccardo Massa," Italy, b.novakovic@campus.unimib.it Abstract. It is expected that innovations in the field of AI, future developments of rapid computers, intercept-proof communications and hyper-sensitive measuring methods will shape our societies in a way where new type of knowledge will be essential for the jobs of the future. AI have the potential to revolutionize education from the ability to analyze large volumes of data quickly and accurately to the AI-supported tutoring and student’s individual learning approaches. But, although rapidly emerging innovations in educational AI are currently among the most promising technological developments which offer many opportunities, they also encounter challenges such as responsible use of trustworthy and ethical AI, data security concerns and how to transform and re-skill workforce for green transition. Human centered AI is not only about technological progress but also about how to align AI with ethical and societal values. It redefines how we interact in an AI-integrated world. As artificial intelligence continue to grow exponentially, the incorporation of new technologies in the education sector can improve interdisciplinary learning environments where teachers can help students in developing their creativity and cognitive abilities. This includes the integration of AI and smart technologies in efficient and creative classes that support both in-class and remote activities. Most of the technologies employed in a smart class rely on AI that empowers the interactive, adaptive, and smart usage of interactive, remote, and mobile computing in physical and/ or virtual environments. Key technologies related to smart classes are based on virtual/augmented/mixed reality, smart environment, computer vision-based surveillance and action (behaviour) recognition, educational chatbots, E-learning platforms, virtual classroom and all screen. AI technologies can be widely applied to enhance personalized and adaptive learning through providing one-to-one tutoring and recommendations for personal learning paths and resources. While intelligent tutoring system are designed to provide customized instruction or feedback to students and promote personalized, adaptive learning based on using AI for instructional content delivery, recommendation of personalized learning path and resource recommendation. Furthermore learning prediction AI enables predicting student learning performance or status in advance through AI algorithms and modeling approaches. This includes the use of AI algorithms and modeling techniques such as educational data mining and machine learning technique to help instructors to adjust the instructional processes by predicting students’ learning performance, as well as the use of AI to predict learning risks and dropout factors in education to help instructors intervene in student learning. As a result a recommendation of personalized learning path is enabled through development of combined genetic algorithms and parliamentary optimization algorithms to create personalized courseware. Another approach is to use recurrent neural network and sequential prediction models to create students ability charts and learning paths based on their submission history. Educational AI can be also used for student behavior detection in order to help teachers to track students’ learning outcomes and to motivate them through monitoring their learning behaviors such as preferred learning materials and self-directed learning performance. It is also possible to use various AI tools to automatically create multiple choice tests through the use of natural language processing techniques. By providing an automated syntactic and structural checking and immediate feedback, automated assessment algorithms can help students to improve their abilities of statement and justification in argumentation. Student behavior detection AI can be also used to analyze and reveal students’ latent behaviors by using data mining and learning analytics to exploit and track students’ learning behaviors, patterns, and characteristics. As a result educational AI will indirectly influence the educational practices and effects by transforming the instructor–student relationships from the instructor-directed to student-centered learning practice. One of the most crucial issues of the future of AI educational systems is ethics in the use of data. It is vital to address the way data are collected and used by those systems in order to avoid the violation of privacy. Having all this in mind the aim of this study is to present an analysis of the applications of AI in education and to provide an initial approach to core terms, ideas, and suitable explanations. Special attention will be paid to the potential of AI in the fields of integrated learning practices, learning predictions, the use of intelligent tutoring systems and student group formation, student behavior detection, automation and the use of educational robots. By organizing numerous workshops on creative use of AI we have found that although AI technologies might have the potential to replace the original role of professor and work as a new subject to individually convey knowledge, they still lack the human ability to convey social emotion, solve critical problems, and implement creative activities. Furthermore some AI technologies like Intelligent tutoring system can lead to the significant improvement of learning performance, (especially for the moderate-level students) improve students’ affective perception, such as attitude, interest, and motivation as well as students’ higher-order thinking, such as problem-solving ability, computational thinking, and self-regulated learning skills. Keywords: Human centered AI, Innovative education, Trustworthy AI References: [1] Nesra Yannier, Scott E. Hudson, Kenneth R. Koedinger, (2020) Active Learning is About More Than Hands-On: A Mixed-Reality AI System to Support STEM Education, International Journal of Artificial Intelligence in Education 30:74–96 [2] Vladimir Pavlović, Nataša Milosavljević, Vera Pavlović, Branislav Vlahović, (2023), Artificial intelligence and digital technologies in digitally-supported university education Mathematics for Human Flourishing in the Time of COVID-19 and Post COVID-19 Published by De Gruyter, 2023, 195-200 [3] Weiqi Xu and Fan Ouyang (2022) The application of AI technologies in STEM education: a systematic review from 2011 to 2021 International Journal of STEM Education 9:59 [4] OECD Digital Education Outlook: Towards an Effective Digital Education Ecosystem, OECD Publishing 2023