Bridging the mentorship divide: how large language models could reshape medical workforce equity

Abstract

Mentorship in predictive large language models (LLMs) presents both opportunities and risks of inequity if these systems are not designed with care. Mentorship quietly shapes medical careers, yet opportunities for it remain unequal. LLMs can analyze student writing and may reveal potential that exams overlook. This predictive capacity could inform fairer mentorship and direct support to students who need it most, helping build a more equitable medical workforce.