German General Practice Trainees Report AI Aids Reasoning but Risks Deskilling

General practice trainees in Germany perceive artificial intelligence as a supportive tool capable of augmenting and accelerating clinical decision-making, while simultaneously identifying potential risks of deskilling and overreliance on automated outputs. According to qualitative research published in BMC Medical Education, participants reported that large language models are already entering day-to-day clinical practice, offering new opportunities to complement reasoning while raising concerns regarding critical appraisal and the preservation of clinical expertise.

The investigation addressed the substantial complexity general practitioners routinely confront when making clinical decisions. While large language models have been proposed to support clinical reasoning and improve access to medical knowledge, their actual impact on decision-making processes and postgraduate medical training has remained unclear. To clarify these dynamics, the study explored how general practice trainees actively utilize these tools and examined their specific educational needs during residency.

Methodologically, the authors designed a qualitative investigation centered on direct trainee experiences. Each interview session was subsequently transcribed verbatim and evaluated through qualitative analysis to capture the nuances of how emerging AI technologies intersect with clinical decision-making.

A total of 17 general practice trainees participated in the interview process. Across the cohort, individual interview lengths ranged from 21 minutes to 63 minutes, providing a focused sample of personal perspectives on technology integration within postgraduate medical education.

In analyzing participant accounts, the researchers observed that trainees primarily viewed artificial intelligence as an aid to augment clinical reasoning and facilitate access to complex medical information. However, this perceived utility was balanced against notable professional reservations. Participants repeatedly highlighted the risk of overreliance on AI-generated information, alongside concerns regarding possible negative consequences for the development and maintenance of independent clinical expertise.

Beyond daily clinical reasoning, the interviewed trainees emphasized a pronounced lack of structured artificial intelligence education within their current training frameworks, articulating an explicit need for formal curricular integration. The authors concluded that because artificial intelligence is already embedded in practice, postgraduate medical programs must incorporate structured competencies focusing on critical appraisal, ethical responsibility, and reflective, patient-centered application to ensure safe and effective clinical integration.

Subject of Research: Artificial intelligence in general practice clinical reasoning and training

Article Title: Artificial intelligence and clinical reasoning – a qualitative study with general practice trainees in Germany

Article References: Dreher, A. C., Sauer, H., Altiner, A., & Schwill, S. (2026). Artificial intelligence and clinical reasoning – a qualitative study with general practice trainees in Germany. BMC Medical Education. https://doi.org/10.1186/s12909-026-10736-3

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10736-3

Keywords: artificial intelligence, large language models, clinical reasoning, clinical decision-making, general practice trainees, postgraduate medical training, medical education, qualitative research, semi-structured interviews, deskilling, overreliance, critical appraisal, ethical responsibility, medical knowledge access, patient-centered care, Germany, general practice