← 返回研究新闻
Academic cover for a state of the field review on AI in higher education
综述证据综述20232026年6月7日· 8 min

Artificial intelligence in higher education: the state of the field

Helen Crompton, Diane Burke

International Journal of Educational Technology in Higher Education

500 字摘要

Academic cover for a state of the field review on AI in higher education

Crompton and Burke update the AI-in-higher-education evidence landscape by reviewing research from 2016 to 2022. The paper is important because it captures the period immediately before and at the edge of the generative AI explosion, showing that higher education AI research was already accelerating before ChatGPT became a public reference point. Using PRISMA principles, the authors identified 138 articles for detailed coding and analysis. The result is a field map that helps readers understand where AI was being studied, who was studying it, and what functions it served.

The review reports rapid growth, especially in 2021 and 2022, with publication numbers rising sharply compared with previous years. It also notes geographical and disciplinary shifts. Earlier reviews had found limited involvement from education departments, but this review reports education as the dominant departmental affiliation among researchers. That makes it a useful companion to Zawacki-Richter et al. 2019, which asked where educators were in AIED research. Crompton and Burke suggest that educator involvement has become more visible, although the field still has gaps.

The paper identifies five major usage categories for AI in higher education: assessment and evaluation, predicting, AI assistant, intelligent tutoring system, and managing student learning. These categories are practical for Research News because they can become tags or filters. They also help AIEDHK separate very different use cases. A predictive model for academic success, a chatbot assistant, an automated assessment tool, and an ITS all require different evidence standards and governance routines.

The review also gives readers a way to see what the field may be over-studying or under-studying. A rapid rise in publications does not automatically mean a mature evidence base. The coding categories reveal where research attention is concentrated and where implementation knowledge may still be thin. For example, if many studies focus on student-facing systems but fewer study instructor workflows, then institutions may lack evidence about teacher workload, professional judgment, and adoption barriers. If language learning is prominent, that may indicate a promising domain, but it should not be generalized to every subject. AIEDHK can use these patterns to plan research coverage that includes gaps, not only popular topics.

The review also highlights who AIED systems are intended to support. Most studies focused on students, with fewer focused on instructors or managers. Language learning was the most common subject domain. Undergraduate students were the most studied group. These findings matter for product strategy: the evidence base may be strongest around student-facing tools and language learning, while teacher-facing and management-facing AI may need more careful development and evaluation. The paper is a current state-of-field entry and should be paired with generative AI work because its search window ends before many ChatGPT-era studies appeared. For AIEDHK, that pairing can show both continuity and rupture in higher education AI research. It also helps readers avoid assuming that today's LLM debate is disconnected from earlier work on assessment, prediction, and tutoring.

相关论文

An East Asian female university student and a South Asian male language tutor compare bilingual economics notes and two AI explanations in a sunlit campus language centre
工具 / 数据集2026
工具 / 数据集 226

Preprint: Students’ access to generative AI did not ensure equally usable support across languages

Ming Li, Qin Xie, Ariunaa Enkhtur, Lilan Chen, Fei Cheng

arXiv preprint (work in progress)

Interviews with 42 undergraduates at universities in China, Japan, and Mongolia explored how students selected AI tools, chose interaction languages, and checked outputs. Participants described different routes from technical access to useful academic support, including switching languages when responses seemed inadequate. This work-in-progress preprint presents situated accounts rather than national comparisons or measured learning gains. It raises practical questions about evaluating an AI service in the languages students actually use.

preprintmultilingual educationAI access
阅读 500 字摘要 →
A South Asian female lecturer and a White male university student examine two differently annotated statistics feedback sheets beside a hand-drawn confidence-interval diagram
期刊论文2026
期刊论文 216

Students’ feedback ratings reflected source credibility without consistently matching expert judgments

Tanya Nazaretsky, Hagit Gabbay, Tanja Käser

Computers and Education: Artificial Intelligence

Researchers compared human and AI feedback on 979 responses from six university STEM courses, then examined ratings from 472 students. Expert analysis found broadly similar pedagogical quality, while students’ judgments were associated with their views of the source’s credibility. The study did not measure feedback uptake or learning gains. For educators, it raises a practical question: how can students justify an evaluation of advice before deciding whether to use it?

feedback literacysource credibilityhigher education
阅读 500 字摘要 →
Academic cover for a systematic review of AI applications in higher education
综述2019
综述 01

Systematic review of research on artificial intelligence applications in higher education: where are the educators?

Olaf Zawacki-Richter, Victoria I. Marin, Melissa Bond, Franziska Gouverneur

International Journal of Educational Technology in Higher Education

A highly cited systematic review mapping AI applications in higher education and asking why educators, pedagogy, and ethical reflection remain underrepresented in the field.

higher educationAI applicationseducator role
阅读 500 字摘要 →