STRENGTHS AND LIMITATIONS OF AUTOMATED ASSESSMENT SYSTEMSSTRENGTHS AND LIMITATIONS OF AUTOMATED ASSESSMENT SYSTEMS

The rapid integration of automated assessment systems (AAS) into higher education has transformed traditional approaches to evaluating student performance, particularly in language and writing instruction. Driven by advances in artificial intelligence and natural language processing, these systems are increasingly employed to provide immediate feedback, ensure consistency in grading, and support large-scale assessment. The present study investigates the strengths and limitations of automated assessment systems within an academic writing context through an experimental research design. A mixed-methods approach was adopted, involving sixty undergraduate students enrolled in an academic writing course at a Kazakhstani university. The experiment consisted of three stages: pre-test, intervention, and post-test. During the intervention phase, students regularly received automated feedback on their written assignments, complemented by instructor guidance. Writing performance was evaluated using standardized criteria, including task achievement, coherence and cohesion, lexical resource, and grammatical accuracy.Quantitative analysis revealed statistically significant improvements across all assessed criteria following the intervention, indicating the effectiveness of automated assessment in supporting writing development. Qualitative data, collected through surveys and reflective responses, demonstrated increased learner autonomy, motivation, and confidence in revising written work. At the same time, the findings identified notable limitations of automated assessment systems, particularly their limited capacity to evaluate content originality, argument depth, and contextual appropriateness. Some participants also reported a tendency to rely excessively on automated feedback. The study concludes that automated assessment systems represent a valuable pedagogical tool when used as a complement to human evaluation rather than a substitute. Integrating automated and instructor-led assessment practices can enhance both the efficiency and educational quality of academic writing instruction.