Agha Ali Raza

Publications/HCI for Development: Health, Gender & Financial Inclusion

Conference · 2026

Empathy Applicability Modeling for General Health Queries

Shan Randhawa, Agha Ali Raza, Kentaro Toyama, Julie Hui, Mustafa Naseem

Findings of ACL 2026, San Diego, CA, USA · CORE A*

Summary

AI systems increasingly answer patients' health questions, but often without the empathy that good clinical communication requires. Rather than judging empathy after a response is written, we model in advance whether a patient's message calls for it. Our Empathy Applicability Framework classifies health queries by whether emotional reactions and interpretations are warranted, based on clinical, contextual, and linguistic cues. We release a benchmark of real patient queries annotated by humans and by an LLM, and train classifiers that outperform heuristic and zero-shot LLM baselines. Remaining challenges include implicit distress and ambiguity about clinical severity.

Cite this paper

Shan Randhawa, Agha Ali Raza, Kentaro Toyama, Julie Hui, Mustafa Naseem. Empathy Applicability Modeling for General Health Queries. Findings of ACL 2026, San Diego, CA, USA.

@inproceedings{randhawa2026empathy,
  author = {Shan Randhawa and Agha Ali Raza and Kentaro Toyama and Julie Hui and Mustafa Naseem},
  title = {{Empathy Applicability Modeling for General Health Queries}},
  booktitle = {Findings of ACL 2026, San Diego, CA, USA},
  year = {2026},
  url = {https://aclanthology.org/2026.findings-acl.1983/}
}

More on this theme

All 10 works on this theme →