Placebo, Nocebo, and Machine Learning: How Generative AI Could Shape Patient Perception in Mental Health Care

Abstract

The emergence of generative artificial intelligence (GenAI) in clinical settings-particularly in health documentation and communication-presents a largely unexplored but potentially transformative force in shaping placebo and nocebo effects. These psychosocial phenomena are especially potent in mental health care, where outcomes are closely tied to patients' expectations, perceived provider competence, and empathy. Drawing on conceptual understanding of placebo and nocebo effects and the latest research, this Viewpoint argues that GenAI may amplify these effects, both positive and negative. Through tone, assurance, and even the rapidity of responses, GenAI-generated text-either co-written with clinicians or peers, or fully automated-could influence patient perceptions in ways that mental health clinicians may not currently fully anticipate. When embedded in clinician notes or patient-facing summaries, AI language may strengthen expectancies that underlie placebo effects, or conversely, heighten nocebo effects through subtle cues, inaccuracies, or potentially via loss of human nuance. This article explores the implications of AI-mediated clinical communication particularly in mental health care, emphasizing the importance of transparency, ethical oversight, and psychosocial awareness as these technologies evolve.

Keywords: ChatGPT; artificial intelligence; ethics; generative language models; health disparities; large language models; nocebo effects; placebo effects.