Inform, Coach, Relate, Listen: Auditing LLM Caregiving Support Roles

2026-05-28

AI Summary

This study addresses the gap in safety evaluation of large language models (LLMs) within informal caregiving contexts, where assessments often rely on generic prompts and overlook the influence of supportive roles on interaction safety. Grounded in social support theory, the authors define four caregiving roles—informing, guiding, empathizing, and listening—and systematically evaluate the interaction risks and support quality of three LLMs using role-conditioned prompting, retrieval-augmented generation controls, large-scale human evaluation, and risk annotation across 5,000 real-world queries from Alzheimer’s caregivers. The work reveals, for the first time, that supportive roles significantly shape risk distributions: information-oriented roles exhibit higher risk yet are perceived by users as more trustworthy and useful. The project releases the first ecologically valid caregiving evaluation dataset, comprising approximately 90,000 risk-annotated model responses.

Abstract

Language models are increasingly being deployed for conversational support in informal caregiving contexts, where interactions often extend beyond information-seeking: caregivers seek emotional reassurance, guidance, and help, while navigating uncertain, relationally complex care decisions. Yet most safety evaluations assess model behavior under generic prompts, leaving a critical question unexamined: does a model's safety profile change with its support role? We study this by operationalizing four expert-reviewed support roles grounded in social support theory: Inform, Coach, Relate, and Listen, and comparing them against two baseline controls: a basic prompting condition and a retrieval-augmented generation (RAG) condition. We evaluate across three language models (GPT-4o-mini, Llama-3.1-8B-Instruct, and MedGemma-1.5-4b-it) on 5,000 real-world queries from online Alzheimer's Disease and Related Dementias (ADRD) communities. We find that the LLM's support role systematically shapes both the prevalence and composition of interactional risks. Furthermore, a human evaluation study reveals a perceived quality--safety tension: more directive, information-oriented roles are rated as more helpful and trustworthy despite exhibiting elevated interactional risk profiles. We release ~90,000 support role-conditioned model responses with risk annotations as an ecologically grounded resource for research on safer LLM-mediated conversational support.

Problem

• LLM safety
• caregiving support roles
• conversational AI
• interactional risks
• social support theory

Innovation

• support roles
• interactional risk
• safety-quality tradeoff
• conversational AI
• ecologically grounded dataset