In medical education, respect for human dignity and for life has long been taught most concretely in the anatomy laboratory; how the students who enter that room are selected, however, has received less attention. This narrative systematic review examines, from the standpoint of character formation, the validity of the multiple mini-interview (MMI) not as a device that measures character as a completed attribute, but as a structured setting in which character-related performances—ethical judgment, interpersonal communication, and professional behavior—are observed, sampled, and cautiously generalized. Six databases were searched on 9 July 2026, identifying 952 records; after de-duplication and screening, 46 full texts were assessed and 45 studies were included (33 primary studies, 11 systematic reviews, and 1 station-construction guide). Methodological quality was appraised with standardized instruments, quantitative findings were synthesized by direction-of-effect vote counting, and the certainty of qualitative findings was formally rated. Three findings emerged. First, MMI stations formed four categories—ethical dilemmas, interpersonal integrity, empathic communication, and public responsibility—each mapped onto specific virtues such as respect for autonomy, justice, honesty, responsibility, compassion, and concern for the common good. Second, the multi-station, multi-rater structure of the MMI was consistently more stable than traditional interviews, although absolute reliability varied widely with the number of stations, rater training, and rubrics; MMI scores discriminated non-cognitive qualities independent of academic achievement and were associated with subsequent clinical and licensing-examination performance, although the licensing evidence rested largely on a single large cohort, and longitudinal evidence for long-term clinical morality remained limited. Third, online formats and stations developed with generative artificial intelligence reached reliability comparable to face-to-face formats, although in the studies compared directly both formats showed low absolute reliability, and they raised concerns about the loss of non-verbal empathy, fairness, data security, value bias, and the need for human verification. As an interpretive proposal rather than an empirically verified claim, the review repositions the MMI as a bridge linking selection to the education of respect for human dignity and life after admission, and to the formation of professional identity. The review was registered in PROSPERO (CRD420261446189).