Medicine as a Human Science: Not Everything that Counts can be Codified

Medicine has always been a human science before becoming biomedical. Here, ‘human science’ refers to the interpretive, relational, and normative dimensions of clinical practice that complement biomedical knowledge. Today, however, clinicians are increasingly overwhelmed by a digital flood of clicks, codes, and administrative tasks that can displace the core values of care. Time—once the heart of clinical presence—has become one of the rarest resources in many healthcare systems. Without time, there can be no listening, shared decision-making, or trust. When designed and governed responsibly, artificial intelligence (AI) may help restore some of this lost time to clinicians. Ambient AI systems that transcribe and summarize patient interactions can reduce documentation fatigue and cognitive overload, enabling clinicians to look up from their screens and reconnect with patients. Yet the ethical value of AI in medicine should not be judged primarily by productivity metrics. Unless the “time dividend” is protected, efficiency gains may be absorbed by financial and systemic pressures that further compress visits. This commentary argues that AI can be a moral technology only if it is implemented with safeguards that preserve professional judgment, patient privacy, and relational care—by co-design with clinicians and patients, transparent governance, and policies that reinvest time saved into the patient–clinician encounter.