AI in Education

AI in education is the application of artificial intelligence to teaching, learning and the administration of education. It spans decades of intelligent tutoring systems, but the arrival of large language models in 2022 made conversational, always-available assistance cheap enough to reach ordinary classrooms — and forced schools, universities and regulators to decide what students may do with it.^[1]

Background

Computer-assisted learning long predates modern AI: systems such as the ITS tradition of the 1990s diagnosed student misconceptions from problem-solving steps and adapted the next question accordingly, and research consistently found one-to-one tutoring among the most effective known interventions. Machine learning then added automatic grading of short answers, speech recognition for language practice, and recommendation engines that sequence practice by difficulty.^[2]

Generative AI changed the ceiling. A general chatbot can act as a Socratic tutor that never tires, explain the same concept five ways, produce unlimited practice problems, give formative feedback on drafts, translate and summarise readings, and draft lesson plans or rubrics for teachers. By 2025–2026, school- and district-level licensing of such tools had become commonplace, with commercial offerings such as Khanmigo alongside general-purpose models used directly by students.^[1][2]

Key applications

• Personalised tutoring — adaptive explanation, hints and practice at the learner's pace, the use case with the strongest evidence base.
• Feedback at scale — formative comments on drafts and problem sets, freeing teacher time for higher-value interaction.
• Assessment support — generating item banks and rubrics, scoring short-answer work, and detecting anomalous submission patterns.
• Accessibility — speech-to-text, translation, read-aloud and simplification for learners with disabilities or limited-language backgrounds.
• Administration — timetable optimisation, admissions triage, and document processing for enrolment and records.^[2][3]

Risks and debate

The central classroom problem is academic integrity: if homework can be generated in seconds, traditional take-home assessment loses diagnostic value, pushing institutions toward in-class writing, process-based assessment and oral defence. Over-reliance is a second concern — studies of AI tutoring find gains when the tool scaffolds reasoning but weaker outcomes when it simply supplies answers. Privacy follows: education tools process data about minors, which engages data-protection law wherever students are identifiable. UNESCO's 2023 guidance recommends a minimum age of 13 for independent use of general-purpose AI in classrooms, teacher training before deployment, and human oversight of any AI used in assessment.^[1][4]

>Key Takeaways

• AI tutoring and feedback are the highest-value uses; evidence supports scaffolding, not answer supply.
• Generative AI has forced a redesign of assessment toward process and oral verification.
• Privacy obligations apply squarely when tools process identifiable data about minors.
• UNESCO's 2023 guidance — age 13 minimum, teacher training, human oversight — remains the reference point for policy.
• In Malaysia, teacher upskilling is well advanced; rural connectivity is the remaining bottleneck.

See Also

• AI in Malaysian Education
• Large Language Models
• AI and Employment
• PDPA Malaysia
• MDEC Malaysia
• AI Ethics