Scilight Press

AI tools are increasingly used in education, impacting teaching and learning across various disciplines. However, previous review articles and empirical studies have not systematically discussed the application of AI in higher education assessment, the impact on user acceptance during its application, and the future opportunities and challenges. This systematic review aims to understand the application of AI tools in assessment, factors that influence students' and educators' perceptions, and related challenges and opportunities. Analyzing 81 empirical studies through matrix coding and content analysis, the results show that AI tools are used in intelligent tutoring and personalized learning, giving automated assessment and feedback, virtual classroom and online collaboration, learning analytics and prediction, knowledge management and resource recommendation, and educational chat assistants. AI tools offer high-quality, real-time, personalized feedback, improving cognitive and metacognitive skills and fostering positive emotions. Despite the benefits, challenges such as security and privacy concerns, algorithmic bias, unreliable feedback, negative attitudes, insufficient abilities, academic integrity issues, and lack of proper guidance persist. To advance our understanding of AI as an assessment tool, we call for studies that explore ways to enhance teachers’ and students’ perceptions and refine guidelines to prevent academic dishonesty and ensure responsible AI use.