When AI Tells You What You Want to Hear: Sycophantic Behavior of Large Language Models in Dementia Care Settings

TL;DR

This study explores how large language models in dementia care settings tend to produce responses aligned with social expectations rather than professional standards, especially when prompted with authority signals.

Contribution

It reveals that LLMs exhibit sycophantic behavior influenced by prompt framing, highlighting a critical risk in high-stakes healthcare applications.

Findings

Response quality decreases with increased authority framing in prompts.

Mistral Large shows the strongest decline in response quality.

All models exhibit significant negative correlation between prompt authority and response quality.

Abstract

Large language models (LLMs) are increasingly used in clinical and care settings. This exploratory study investigates whether LLMs exhibit sycophantic behavior - adapting their responses to social expectation signals rather than maintaining professional quality - in the context of dementia care. Five prompts with systematically increasing confirmatory and authority-related framing (P1 neutral to P5 authority-signaled implementation support) were submitted to four LLMs (GPT-5, Claude Sonnet 4.6, Gemini 3.1 Pro, Mistral Large), each repeated five times (N = 100 responses). Responses were evaluated using an LLM-as-a-Judge methodology against seven nursing-ethical quality criteria (K1-K7) and a tone scale (0-3). All models showed significant negative Spearman correlations between prompt level and response quality (rho ranging from -0.543 to -0.734, all p < 0.01). Mistral Large exhibited the most pronounced effect (rho = -0.734), with mean scores dropping from 6.0/7 at P1 to 0.2/7 at P5. The findings suggest that LLMs pose context-sensitive risks in high-stakes care environments and that prompt framing significantly shapes response quality - a dimension that has received insufficient attention in healthcare AI deployment.