As Artificial intelligence (AI) rapidly transforms the health-care landscape, it raises serious concerns over its implementation – and whether the focus will be on efficiency or human-centred care.
An efficiency-driven AI approach in health care focuses on maximizing throughput. By automating administrative tasks, AI may help clinicians recover time and be framed as a solution to patient backlogs and resource shortages. But prioritizing efficiency runs the risk of organizations using AI to justify increasing patient volumes, reducing health care to a transactional process that erodes the human connection at the core of patient-physician relationships.
In contrast, a connection-driven approach seeks to leverage AI to enhance human interactions and deepen the art of medicine. As Eric Topol suggests, “The greatest opportunity offered by AI is not reducing errors or workloads, or even curing cancer: it is the opportunity to restore the precious and time-honoured connection and trust” between patients and doctors. In this model, AI is not merely a tool for productivity but becomes a partner in fostering meaningful patient experiences.
Despite the increasing complexity of care and rising administrative demands, the time allocated for patient visits has remained largely unchanged. There are many that argue primary care simply cannot be delivered effectively within current time constraints. One study estimates that providing all guideline-based care for an average family medicine patient roster would require an impossible 27 hours per day.
Health-care systems often prioritize what is easiest to measure – such as patient volume or procedural efficiency. Intangible qualities that characterize “care” such as trust, empathy and human connection are poorly measured. We must begin the task of resisting the pull of efficiency-driven care today rather than scrambling in the future at the prospect of a health-care system devoid of humanity.
Inevitably, AI will become a third party in the patient-doctor relationship.
A human-centred AI implementation philosophy should focus on the experience of patients and clinicians as its foundation. It should ensure end-users of the technology are actively involved in the design, study and maintenance of tools and programs. Human-centred evaluations of AI tools should be robust and capture what is most important to clinicians and patients through qualitative research methods.
Inevitably, AI will become a third party in the patient-doctor relationship – one that both physicians and patients must learn to trust. Yet, the need for human-to-human conversations will not disappear. In fact, experts argue that doctors will have to develop even greater skills in interpreting and communicating as they will need to communicate AI-generated insights to engage in shared decision-making with their patients.
AI tools, no matter how powerful, cannot substitute for the therapeutic relationship. AI may be able synthesize vast datasets, but it will never hold a patient’s hand through uncertainty or guide them in deeply personal health-care decisions. It will never replace the profound care and drive that clinicians provide to protect and heal their patients. AI may be trained to say very comforting things, but it can’t draw on personal experiences to relate to families grieving over the loss of a loved one.
AI is neither inherently good nor bad – it is a tool. The responsibility lies with us as clinicians and health-care leaders to ensure its adoption strengthens, rather than diminishes, the humanity of care.
