Saved by the Bot! A Comparative Analysis of Human vs. AI Tutors on Formative Assessment Performance of Computer Science Students in Higher Education

Generative AI in Higher Education has become a prevalent point of discussion for practitioners and researchers alike. Current literature looks at such technologies either on negative aspects such as plagiarism and cheating (Bittle, 2025), or on the positive uses of this technology for assessment (Zhang, 2024), with little discussion focusing on the integration of such technologies as digital tutors. In this study, we investigate the performance different in Computer Science students results in a formative assessment activity between groups using a human tutor and another using AI tutors. The AI tutors in this study were constructed using Microsoft Copilot and were prompted using the CRTRF framework (McCormack, 2023). Findings suggest that there are negligible differences between these groups in post-test results, suggesting that AI tutors are just as effective as human tutors, and can be effectively leveraged and scaled to assist in resolving Blooms “Two Sigma Problem” (Bloom, 1984).