Empathy as feeling: reflections on the meaning of clinical empathy in the age of artificial intelligence

Abstract: The importance of expressing empathy in clinical practice is well-recognized by medical professionals, with beneficial effects including strengthening the relationship between clinicians and patients, building trust, and facilitating information exchange. Interestingly, artificial intelligence has been shown to convincingly convey empathy, at least in text. A concerning corollary of this increased access and ease of generating empathetic responses to patients is that empathy may function, in essence, to pacify patients and disincentivize clinicians from advocating for their patients more vigorously. However, despite an abundance of literature aiming to define clinical empathy, consensus does not exist on what clinical empathy is, with most definitions describing empathy as a narrow set of behaviors performed by clinicians in response to specific emotional cues expressed by patients. This presentation aims to make an argument for conceptualizing empathy in medicine as a state of sharing the feelings of patients. In doing so, the definition of clinical empathy moves away from its existing behavioralist account, and instead situates empathy as a mental state that gives rise to compassionate behaviors. Such a conception allows clinicians to not only recognize the difficulties that their patients may be encountering; it allows them to be affected by those experiences as well. This affective aspect of empathy as experiencing patients’ hardships with them can provide benefits that cannot be realized with artificial intelligence or with existing definitions of clinical empathy, as it has the ability to confer powerful, personalized motivations for clinicians to address their patients’ needs.