How AI researchers and clinicians navigate the language barrier in healthcare innovation

Teams developing AI tools for healthcare face a fundamental challenge: bridging the knowledge gap between machine learning experts, clinicians, and patients. A new qualitative study based on interviews with early-stage AI research consortia in the UK uncovers how teams actually navigate this complex landscape.

The translation problem#

Researchers working across disciplines must continually adapt both their communication methods and the substance of what they share. The study found that participants frequently had to recalibrate how they explained technical concepts depending on whether they were speaking with data scientists, clinicians, or patients. This isn't merely about simplifying jargon—it reflects deeper differences in how these groups understand problems and solutions.

The research framework treats AI development as inherently sociotechnical: the technical algorithms, data pipelines, and models cannot be separated from the social interactions, institutional contexts, and human perspectives that shape them. This perspective helps explain why adding AI to established healthcare workflows introduces novel coordination challenges.

Key findings#

The study of 13 semi-structured interviews identified several critical areas shaping collaboration:

• Novelty and unfamiliarity with AI as a research domain across all parties
• Knowledge-sharing practices that must flex based on audience expertise
• Meaningful inclusion of clinician and patient perspectives in model development

Researchers reported needing to invest substantial effort in translating between technical and clinical languages, suggesting that AI projects cannot succeed through technical excellence alone.

Why it matters#

AI-powered healthcare promises to augment clinical decision-making and improve patient outcomes. But realizing that promise depends on teams that can genuinely integrate diverse expertise. This study suggests that interdisciplinary friction in AI healthcare research is not a minor coordination problem to be solved by better project management—it reflects fundamental differences in how professionals across disciplines conceptualize the problem and the solution. Understanding these tensions is a prerequisite for building tools that clinicians will actually trust and patients will accept.