Empathy in AI for Health and Care Settings-Definition, Expression, and Measurement: Protocol for a Scoping Review

Area of Science:

• Health Informatics
• Human-Computer Interaction
• Artificial Intelligence Ethics

Background:

• The concept of empathy in artificial intelligence (AI) systems, particularly in healthcare, is inconsistently defined and applied.
• Lack of clarity in defining and assessing "empathic AI" hinders understanding of user experience and engagement with AI-mediated support.

Purpose of the Study:

• To systematically map the definitions, operationalization, and evaluation of empathy in peer-reviewed AI research within health and care settings.
• To identify and describe interactional design features of AI systems that are perceived as more empathic.

Main Methods:

• A scoping review protocol adhering to Joanna Briggs Institute guidance and PRISMA-ScR standards.
• Comprehensive searches across multiple databases (PubMed, Embase, PsycInfo, CINAHL, Scopus, IEEE Xplore, ACM Digital Library).
• Thematic synthesis of extracted data on AI system types, empathy definitions, design features, and evaluation methods.

Main Results:

• The review will summarize current definitions of empathy in AI research for health and care.
• It will provide a categorized list of empathic interactional behaviors and design features in AI systems.
• An overview of how empathy is measured across studies will be presented, highlighting features in highly-rated systems.

Conclusions:

• This review will clarify the meaning of "AI empathy" in healthcare contexts.
• Findings will identify common system features used to develop empathic AI.
• The results aim to inform the design of more effective and user-centered empathic AI systems.