Clinical Research Associate Baylor College of Medicine, Center for Medical Ethics and Health Policy Houston, Texas
Abstract: Computer perception (CP)—the automated detection, interpretation, and analysis of patient data often collected outside clinical settings (e.g., via wearables)—has emerged as a cornerstone of precision medicine, promising more objective, continuous, and personalized clinical assessments. However, ethicists have raised concerns about its compatibility with humanized care, a patient-centered approach emphasizing empathy, shared decision-making, respect for dignity, and holistic support. To better understand stakeholders’ views on the utility and acceptability of CP in clinical care, we conducted an NCATS-funded study involving interviews (n=100) with clinicians, developers, ethics/legal/philosophy scholars, patients, and caregivers. Our findings highlight several key challenges to humanized care introduced by CP: (1) Overemphasis on Quantifiable Indicators: CP’s reliance on measurable functioning risks overshadowing subjective patient experiences, diminishing human connection, neglecting context-dependent phenomena that remain difficult to computationalize, and overlooking emotional and sociocultural dimensions essential for empathic, dignified care; (2) Threats to Privacy and Autonomy: CP can undermine trust in the therapeutic alliance by reducing patient agency over what they disclose and heightening fears of misuse or discrimination; and (3) Counterproductive Patient Reactions: CP feedback may provoke fixation on metrics, self-censorship, guilt, anxiety, and other responses that undermine clinical progress. Despite these risks, we argue that CP can also enhance humanized care, if integrated strategically. We offer several concrete suggestions, including participatory approaches to co-develop shared “roadmaps” for deploying CP that would prioritize measuring aspects of health most meaningful to patients, enable selective disclosure of personal data, and actively seek to bridge—rather than separate—experiential and data-driven insights.
Keywords: artificial intelligence, digital phenotyping, humanized care
Learning Objectives:
After participating in this conference, attendees should be able to:
understand key stakeholder perspectives on the most pressing ethical and practical concerns and risks of integrating computer perception technologies into clinical care
discuss how CP technologies may contribute to the dehumanization of patient care and how it may be implemented to promote humanistic, personalized care
identify concrete, participatory methods for incorporating CP in ways that protect patient agency, respect individual dignity, and foster a balanced blend of data-driven and experiential insights.
.png)