Abstract:

Artificial intelligence has evolved from early intelligent tutoring systems to contemporary large language models that are reshaping how universities teach, assess, and conduct research. This paper presents a comprehensive review of AI in academia, synthesizing more than seven decades of development from PLATO and early intelligent tutors to current generative AI applications in higher education. It draws on a systematic analysis of over 400 peer‑reviewed studies to examine eight key dimensions: historical foundations, current applications in teaching and learning, documented research gaps, tensions between expectations and realities, productivity effects, institutional implementation patterns, emerging challenges, and future directions. The results indicate that AI‑enhanced systems can produce substantial improvements in student learning outcomes, particularly when used as intelligent tutoring and feedback tools, yet the evidence base remains uneven across disciplines and educational levels. At the same time, the rapid diffusion of generative AI raises persistent concerns about academic integrity, transparency, and equitable access for students and faculty. The review identifies major gaps in rigorous empirical evaluation, faculty readiness, and reproducible research on AI’s long‑term educational impact. Based on these findings, the paper proposes recommendations for institutional policy, curriculum design, and faculty development aimed at integrating AI tools in ways that enhance learning, support ethical academic practice, and sustain research productivity in higher education.

Keywords:

AI in education, intelligent tutoring systems, generative AI, learning outcomes, academic integrity, research productivity, higher education, large language models, ChatGPT, research gaps.