Validating an AI-assisted comentoring model for identifying at-risk students and for academic mentoring: a study...

Area of Science:

• Medical Education
• Artificial Intelligence in Education
• Pedagogical Assessment

Background:

• Academic mentoring is crucial for student progress and early risk identification.
• Traditional mentoring relies on face-to-face interactions, limiting scalability.
• Advancements in AI offer opportunities for AI-assisted mentoring, but existing models are complex.

Purpose of the Study:

• To validate a simplified AI comentor model for efficient at-risk student identification.
• To support continuous academic monitoring with a focus on pedagogy.
• To evaluate the feasibility, acceptability, and analytic agreement of an AI-assisted assessment framework in medical education.

Main Methods:

• Prospective mixed-methods pilot design with ~40 undergraduate medical students and faculty.
• AI component utilized unsupervised machine learning for student competency profiling.
• Feasibility and acceptability assessed via surveys, usage metrics, and qualitative feedback; analytic agreement evaluated using statistical measures.

Main Results:

• The AI comentor model demonstrated efficiency in identifying at-risk students.
• Feasibility and acceptability of the AI-assisted framework were confirmed through mixed-methods evaluation.
• Analytic agreement between AI-derived profiles and faculty assessments showed promising results.

Conclusions:

• A simplified AI comentor model can effectively support academic monitoring and intervention.
• AI-assisted assessment frameworks are feasible and acceptable in medical education.
• This approach has the potential to alleviate mentor workload and enhance student performance and retention.