This paper proposes a human-centered approach to using AI in higher education that promotes equitable access to knowledge while preserving privacy and ethics. Using third-generation Activity Theory, it examines the interaction between three activity systems in higher education: AI teachers, human teachers, and students. The multi-role design recognizes the roles of different actors and AI-based tools, highlighting the importance of understanding how these systems interact and influence each other to design ethical educational technology solutions. By constructing an interaction space between these systems, the paper gains a deeper understanding of the dynamics of teaching-learning in higher education. Based on this analysis, the Ethical AI in Education (EAIED) multi-platform is designed, integrating recent developments in AI in education systems, pedagogical strategies, and ethical guidelines. EAIED emphasizes data privacy, ethics, and interoperability between learning systems. This approach offers interactive, personalized, and equitable teaching-learning experiences that lead to developing critical thinking, higher-order skills, improving engagement, communication, and expansive learning. The integration of AI in higher education has the potential to revolutionize the field by introducing automation, intelligent tutoring systems, personalized programs, adaptive learning, and decision-making. Research has identified key areas of AI applications in teaching and learning, including profiling and prediction, intelligent tutoring systems, assessment and evaluation, and adaptive systems and personalization. Recent studies have found that AI is used for assessment/evaluation, predicting, AI assistant, ITS, and managing student learning. ITS systems use AI to implement intelligent tutor systems in learning platforms for all learning models. Assessment tools, like online-proctoring systems, benefit from AI and machine learning techniques, and provide invigilator programs that supervise students during online exams. MOOCs, as an online learning platform, have recently begun using AI tools regularly, such as online-proctoring. SNSs can supplement traditional LMSs by helping students meet pedagogical objectives. SNSs augmented with education-plugins and/or enhanced with pedagogical strategies could help in communication between students, knowledge creation, knowledge construction, and argumentation skills. Learners can develop higher-order skills and improve their engagement, communication, and learning. This is particularly true when implementing teaching strategies that use SNSs, such as Argumentative Knowledge Construction, Automated Social Learning, and Reciprocal Peer Tutoring. Applying AI to big data gathered in SNSs could generate valuable information of interest to students.This paper proposes a human-centered approach to using AI in higher education that promotes equitable access to knowledge while preserving privacy and ethics. Using third-generation Activity Theory, it examines the interaction between three activity systems in higher education: AI teachers, human teachers, and students. The multi-role design recognizes the roles of different actors and AI-based tools, highlighting the importance of understanding how these systems interact and influence each other to design ethical educational technology solutions. By constructing an interaction space between these systems, the paper gains a deeper understanding of the dynamics of teaching-learning in higher education. Based on this analysis, the Ethical AI in Education (EAIED) multi-platform is designed, integrating recent developments in AI in education systems, pedagogical strategies, and ethical guidelines. EAIED emphasizes data privacy, ethics, and interoperability between learning systems. This approach offers interactive, personalized, and equitable teaching-learning experiences that lead to developing critical thinking, higher-order skills, improving engagement, communication, and expansive learning. The integration of AI in higher education has the potential to revolutionize the field by introducing automation, intelligent tutoring systems, personalized programs, adaptive learning, and decision-making. Research has identified key areas of AI applications in teaching and learning, including profiling and prediction, intelligent tutoring systems, assessment and evaluation, and adaptive systems and personalization. Recent studies have found that AI is used for assessment/evaluation, predicting, AI assistant, ITS, and managing student learning. ITS systems use AI to implement intelligent tutor systems in learning platforms for all learning models. Assessment tools, like online-proctoring systems, benefit from AI and machine learning techniques, and provide invigilator programs that supervise students during online exams. MOOCs, as an online learning platform, have recently begun using AI tools regularly, such as online-proctoring. SNSs can supplement traditional LMSs by helping students meet pedagogical objectives. SNSs augmented with education-plugins and/or enhanced with pedagogical strategies could help in communication between students, knowledge creation, knowledge construction, and argumentation skills. Learners can develop higher-order skills and improve their engagement, communication, and learning. This is particularly true when implementing teaching strategies that use SNSs, such as Argumentative Knowledge Construction, Automated Social Learning, and Reciprocal Peer Tutoring. Applying AI to big data gathered in SNSs could generate valuable information of interest to students.