Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Toward LLM-Powered Social Robots for Supporting Sensitive Disclosures of Stigmatized Health Conditions

Alemitu Bezabih, Shadi Nourriz, C. Estelle Smith

TL;DR

This paper examines how LLM-backed social robots could support sensitive health disclosures, using HIV status as a case study. It adopts a four-step clinical model of disclosure to identify opportunities, design considerations, and risks, and proposes a hybrid architectural approach that combines structured interactions with LLM-driven natural language for select steps. The authors discuss safety, privacy, and ethical concerns, and outline an experimental design to compare disembodied versus embodied robot assistance in an HIV clinic setting. The work aims to inform responsible deployment by highlighting feasible task allocation, privacy protections, and evaluation against standard care to quantify benefits and harms in real-world clinical workflows.

Abstract

Disclosing sensitive health conditions offers significant benefits at both individual and societal levels. However, patients often face challenges due to concerns about stigma. The use of social robots and chatbots to support sensitive disclosures is gaining traction, especially with the emergence of LLM models. Yet, numerous technical, ethical, privacy, safety, efficacy, and reporting concerns must be carefully addressed in this context. In this position paper, we focus on the example of HIV status disclosure, examining key opportunities, technical considerations, and risks associated with LLM-backed social robotics.

Toward LLM-Powered Social Robots for Supporting Sensitive Disclosures of Stigmatized Health Conditions

TL;DR

This paper examines how LLM-backed social robots could support sensitive health disclosures, using HIV status as a case study. It adopts a four-step clinical model of disclosure to identify opportunities, design considerations, and risks, and proposes a hybrid architectural approach that combines structured interactions with LLM-driven natural language for select steps. The authors discuss safety, privacy, and ethical concerns, and outline an experimental design to compare disembodied versus embodied robot assistance in an HIV clinic setting. The work aims to inform responsible deployment by highlighting feasible task allocation, privacy protections, and evaluation against standard care to quantify benefits and harms in real-world clinical workflows.

Abstract

Disclosing sensitive health conditions offers significant benefits at both individual and societal levels. However, patients often face challenges due to concerns about stigma. The use of social robots and chatbots to support sensitive disclosures is gaining traction, especially with the emergence of LLM models. Yet, numerous technical, ethical, privacy, safety, efficacy, and reporting concerns must be carefully addressed in this context. In this position paper, we focus on the example of HIV status disclosure, examining key opportunities, technical considerations, and risks associated with LLM-backed social robotics.
Paper Structure (15 sections, 2 figures)

This paper contains 15 sections, 2 figures.

Table of Contents

  1. Introduction
  2. Prior Work in Sensitive Disclosures
  3. Background: The 4-Step Clinical Model of Disclosure
  4. Key Opportunities for Integration of Social Robots in HIV Clinics
  5. An Example Scenario
  6. Technical & Design Considerations
  7. Which steps are [in]appropriate for a social robot to manage?
  8. How might LLMs be technically integrated?
  9. Reporting concerns
  10. Risks & Ethics in HIV Disclosure
  11. Safety and efficacy
  12. Privacy & HIPAA
  13. Ethical and social concerns
  14. Proposed Experimental Design
  15. Conclusion

Figures (2)

  • Figure 1: A four-step self-disclosure model used in clinical practice, for example by the California Prevention Training Center in the UCSF Bixby Center for Global Reproductive Health california_ptc_self-disclosure_2024
  • Figure 2: A storyboard depicting a scenario in which a user discovers that he is HIV-positive during a clinical visit. The clinician recommends that he visits with a social robot to practice disclosing his HIV-positive status to important social connections.