Defining the ideal artificial intelligence tool by and for GPs in the Nord-Pas-de-Calais region in France: a qualitative study

Background Artificial intelligence (AI) is increasingly discussed as a potential tool in primary care, yet its acceptability and perceived role among GPs remain insufficiently explored, particularly regarding clinical practice and the doctor–patient relationship. Aim To explore GPs’ perspectives in Nord-Pas-de-Calais on the potential role, acceptability, and conditions of use of AI tools in clinical practice. Design & setting Qualitative study using inductive thematic analysis, conducted in the Nord-Pas-de-Calais region between February 2024 and May 2024. Method Eleven GPs were recruited using purposive maximum-variation sampling. Semi-structured interviews explored anticipated or hypothetical uses of AI, as participants did not have routine prior experience with AI tools in general practice. Interviews were continued until data adequacy was reached and confirmed by two consolidation interviews. Results Participants described AI as potentially useful if it remained simple, rapid, integrated into existing clinical software, and under clinician control. They anticipated benefits for documentation, alerts, decision support, continuing learning, and workload relief, but expressed concerns about reliability, data governance, medico-legal responsibility, cognitive over-reliance, and possible effects on the doctor–patient relationship. Conclusion GPs described a pragmatic and conditional openness to AI. Successful implementation in general practice will depend less on technological performance alone than on usability, clinical validation, transparent governance, software interoperability, and the capacity of AI to support — rather than replace — professional judgement and the doctor–patient relationship.