Application of Cognitive Empathy Elements Into AI Chatbots: An Interview Study Exploring Patient-Physician Interaction

Among different components of empathy, cognitive empathy is a necessary facet to understand a patient’s perspectives tapping into their thoughts and reasonings. With the rise of AI chatbots within healthcare settings, it has become important to mirror these cognitive empathetic elements within these chatbots too. Symptom assessment AI chatbots can be valuable tools for initial screening or triage in healthcare settings. This paper seeks to understand the cognitive empathy elements present in patient-physician interactions and offers recommendations for integrating these elements into AI chatbots. This represents a novel exploration of cognitive empathy within the patient-physician context and the potential for incorporating it into patient-AI interactions. We interviewed 10 first-time mothers to explore possible communication patterns for establishing cognitive empathy between patient-physician. We identified incidents with critical interactions where the physician or midwife could develop a shared understanding or common ground with their patients even if they were not aligned at the beginning of the incidents. Six themes emerged from the interviews: (1) Shared Information, (2) Shared Decision-making, (3) Shared Sensemaking, (4) Shared Goals, (5) Communication about Outcomes, and (6) Tailoring to Circumstances. Based on the themes and the interactions extracted from the interviews, we developed design recommendations for symptom assessment/diagnostic reasoning AI chatbots. The design recommendations we provided in this paper offer guidance in developing cognitively more empathic AI chatbots as they are becoming an important part of building robust digital health solutions. Cognitive empathetic elements within these chatbots would allow the users or patients to make improved and empowered choices, decisions, and outcomes through optimal contribution from them with a complete information base.