Learning is most effective when it occurs through personalized experiences that improve skills, knowledge, insights, and understanding of learners. In last decades, personalized learning has emerged as an efficient strategy to provide more meaningful education in accordance with the specific needs of each student. For instance, this personalization could be implemented in instructional designs that consider aspects such as learner prior knowledge, motivations, skills, expertise, preferences, among others. During the process of implementing planned instruction, we expect that environments defined for personalized learning (e.g., digital environments, classrooms, etc.) are capable of providing a responsive instructional experience that enables greater engagement into a given learning task. Such responsive instructional experiences are facilitated by artificial intelligence-based systems. Such systems enable actions performed by intelligent agents that require a larger understanding of learner`s profile through the processing of learners' characteristics. In this case, personalized learning models based on AI could support systems to better understand the individual needs and goals of students, thus making their understanding and investigation essential to learn how personalization occurs in AI systems. In this sense, a systematic mapping review can provide a broad analysis of existing literature by enabling a comprehensive overview of current trends and identifying lacking areas in research. Therefore, this work employs a systematic mapping review to identify issues related to the instructional design of AI-driven personalized systems into educational contexts. Basically, this study aims to systematically map existing AI-driven personalized learning systems to identify key design considerations and inform future educational practices. We define a protocol that outlines the elements considered in this review, such as objectives, research questions, sources, search strategies, and selection criteria. The findings highlight the topics, solutions (approaches, techniques, mechanisms, or methods), tools, and subdomains commonly utilized in the design of personalized educational systems, as well as potential areas for future research. We interpret the findings of this review as an opportunity for educators, engineers, and computer scientists to conduct research on themes related to the instructional design of AI-powered personalization into educational scenarios.

Keywords: Personalization, Personalized Learning, AI-powered personalization systems, Higher Education.