Artificial Intelligence in Education: A Review

Chen, Chen, and Lin provide a broad review of artificial intelligence in education with a practical focus on how AI affects administration, instruction, and learning. The paper is widely cited because it offers a general map for readers who need an accessible overview rather than a narrow technical treatment. It defines AI broadly as systems with human-like cognitive capacities such as learning, adaptation, and decision-making, then follows how these capacities appear in educational institutions through computer-based systems, online platforms, embedded technologies, humanoid robots, chatbots, and adaptive learning tools.

The review argues that AI has been adopted across multiple educational functions. In administration, AI can help institutions handle grading, assignment review, scheduling, student support, and management tasks more efficiently. In instruction, AI can support teachers through content delivery, tutoring, automated feedback, and learning analytics. In learning, AI can personalize materials and pathways based on learner needs, performance, or progress. The paper therefore treats AIED as a set of institutional and pedagogical transformations rather than a single technology.

One reason this paper works well for Research News is that it introduces non-specialist readers to the breadth of the field. It shows that AI in education is not limited to chatbots or intelligent tutors. It includes data mining, adaptive instruction, online intelligent systems, decision support, automation, and feedback. At the same time, the review has limitations that should be visible in editorial framing. It is a broad narrative review, not a deeply critical systematic review. Its categories are useful for orientation, but readers should not treat it as the final evidence base for effectiveness or safety.

The review is also useful because it shows how quickly the meaning of AI in education shifts across audiences. For administrators, AI may mean efficiency, prediction, or student support workflows. For teachers, it may mean instructional assistance, feedback, or analytics. For learners, it may mean personalized content, dialogue, or assessment. AIEDHK can use this distinction to avoid one-size-fits-all claims. The same algorithmic technique can have different value and risk depending on whether it is used for low-stakes practice, high-stakes grading, institutional decision-making, or learner self-study. That framing is particularly important for Hong Kong's multilingual education context, where translation, feedback, and access functions may look attractive but still require validation against curriculum, language level, and equity.

For AIEDHK, the paper can function as a landscape piece. It supports the Research News page's need to explain why AI education matters across teaching, learning, and administration. It also gives a vocabulary for classifying future entries: AI for institutional management, AI for teacher support, AI for student learning, and AI for personalization. The practical implication is that AIED products should be evaluated by the educational function they serve, because a chatbot used for administration should be assessed differently from a tutor used for formative feedback. That functional lens keeps discussion practical for readers who are comparing research, products, and school needs.