JMIR Publications

ABSTRACT

Background:

Healthcare systems face rising demand and persistent staff shortages, intensifying pressure on the quality and sustainability of care. Artificial intelligence (AI) is increasingly introduced to improve efficiency and decision support across clinical domains. While these tools promise operational gains, they can also reconfigure how physicians work, make judgments, and interact with patients, all elements of physicians’ craftmanship. However, most research emphasises technical performance rather than AI’s broader implications for physicians’ craftmanship.

Objective:

To explore how physicians define craftsmanship in medicine and how they perceive AI to influence its professional and personal dimensions, with the goal of deriving practical principles for responsible AI design and implementation in hospital care.

Methods:

We conducted a qualitative, exploratory study in two phases (December 2024–September 2025). Phase 1 involved semi-structured interviews with 20 physicians from different hospital types and diverse specialties within the Netherlands. Phase 2 comprised two focus groups with physicians, medical specialists in training, policymakers, and AI developers during a national symposium, using an interactive, persona based design to co create practical design principles.

Results:

Physicians described craftsmanship as their commitment to deliver the best possible care through human judgment, empathy, and contextual understanding. Perceived AI effects clustered in two areas: 1. Professional dimensions: AI was seen to support workflow efficiency, documentation, data integration, and aspects of analytical reasoning, potentially freeing time for patient contact and reflection. Conditions for adoption included human-in-the-loop oversight, explainability, traceability, and AI literacy. 2. Personal dimensions: empathy, contextual interpretation, and ethical judgment were viewed as inherently human and resistant to substitution. Concerns centred on de-skilling, less room for independent judgment, and threats to professional autonomy. Specialties differed in how they framed tasks and AI’s role, reflecting their specific clinical contexts, but all shared the same core aim of delivering high-quality care. Focus groups yielded six main design principles: start from real clinical needs; design AI as supportive, not intrusive; safeguard autonomy and trust; design for contextual diversity; strengthen the role of professional groups; and use AI as a mirror for reflection and learning.

Conclusions:

AI seems to affect the conditions of professional craftmanship, and thereby indirectly the personal dimensions of it. This should be considered in design and implementation, while recognising that continued interaction with AI may gradually reshape what craftsmanship itself comes to mean.