Beyond Algorithms: Protecting the Reasoning Skills of Future Clinicians in an AI-driven Era

Abstract

It’s no longer a fantasy to have an AI-empowered junior doctor who can use algorithms to diagnose clinical conditions. The influence on clinicians’ thinking as AI enters clinical routines, from radiology to emergency triage, is profound. AI can bring precision and efficiency, but misapplication of the technology could degrade the fundamental cognitive skills that make up clinical expertise. Today the question is not whether to use AI in clinical practice, but how to use it effectively, preserving the instinctive experience-based judgement of clinicians, while benefiting from the technological support. As we become more reliant on artificial intelligence it is important to consider how this may influence clinicians’ thinking in years to come. Will they carefully consider each case or will AI be their sole diagnostic tool?

AI is good at pattern recognition, data synthesis and probabilistic reasoning. Technologies like IBM Watson, DeepMind’s AlphaFold and clinical decision support systems have demonstrated impressive results in diagnosing complex diseases and predicting outcomes.¹ In high-stakes settings, AI can help reduce diagnostic errors, streamline processes and identify subtle differences that may be missed by humans. Yet, this capability comes with a cognitive cost. Frequent dependence on artificial intelligence may result in cognitive offloading the practice of delegating mental tasks to external systems, thereby diminishing internal reasoning capabilities.² Healthcare professionals might start to place greater trust in algorithms than in their own judgment, particularly when faced with time constraints or uncertainty. As a result, they may have opportunities to get involved in problem-solving, which may further reduce their reasoning abilities.³