The growing integration of artificial intelligence in programming education has led to the adoption of AI-assisted grading and feedback systems in large-enrollment courses. This study examines students’ perceptions of Gradify, an AI-assisted autograding platform with optional AI-generated feedback, deployed in a second-year undergraduate course (Advanced Object-Oriented Programming). An anonymous end-of-term survey was administered to the enrolled students. Findings indicate that students primarily used AI-generated feedback as an early-stage debugging aid to interpret failed test cases and refine implementations. While many reported improved debugging efficiency and solution quality, perceptions of its contribution to deeper conceptual understanding were mixed. Trust in AI feedback was conditional and influenced by concerns about occasional inaccuracies. Students were mostly comfortable with automated grading when paired with human oversight, particularly for open-ended assessments. Overall, AI-assisted grading was viewed as a supportive tool rather than a replacement for instructor judgment.
Awais Aziz, Marzieh Ahmadzadeh· Proceedings of the Canadian...· 0 citations
Delivering high-quality formative feedback in large introductory programming courses remains challenging despite the widespread use of automated grading systems. Although large language models (LLMs) can generate natural-language explanations grounded in student code and test outcomes, their diagnostic reliability and pedagogical suitability require systematic evaluation. This study presents a comparative analysis of six LLMs for generating formative feedback on introductory Java programs containing predefined defects under controlled conditions. Model outputs were evaluated across key dimensions of feedback quality and performance. Results reveal substantial cross-model variation, particularly in multi-defect scenarios. More precise fault localization was associated with higher diagnostic accuracy and completeness, but also with more frequent instructional-constraint violations. No single model consistently outperformed others across all criteria, and response latency did not reliably correspond to diagnostic quality. These findings underscore the context-dependent nature of LLM-generated formative programming feedback.
Melina Najimi, Saba Yazdani, Marzieh Ahmadzadeh· Proceedings of the Canadian...· 0 citations
