Adaptive LLM Agents: Toward Personalized Empathetic Care

TL;DR

This paper introduces an adaptive framework for mental-health chatbots that personalizes interactions based on a user's psychological state, using large language models and continuous feedback to improve therapeutic engagement.

Contribution

It presents a novel adaptive architecture with specialized agents based on illness acceptance levels, integrating clinical insights and speculative design fiction to explore future mental-health support systems.

Findings

Demonstrates feasibility of personalized LLM agents for mental health care.

Shows how continuous feedback can tailor therapeutic interactions.

Explores societal impacts through narrative scenarios.

Abstract

Current mental-health conversational systems are usually based on fixed, generic dialogue patterns. This paper proposes an adaptive framework based on large language models that aims to personalize therapeutic interaction according to a user's psychological state, quantified with the Acceptance of Illness Scale (AIS). The framework defines three specialized agents, L, M, and H, each linked to a different level of illness acceptance, and adjusts conversational behavior over time using continuous feedback signals. The AIS-stratified architecture is treated as a diegetic prototype placed in a plausible near-future setting and examined through the method of design fiction. By embedding the architecture in narrative scenarios, the study explores how such agents might influence access to care and therapeutic relationship. The goal is to show how clinically informed personalization, technical feasibility, and speculative scenario analysis can together inform the responsible design of LLM-based companions for mental-health support.