Year: 2026
Type: conference-paper
Abstract: Large Language Models (LLMs) often struggle with persuasion in high-stakes scenarios.People's individual personalities and concerns require tailored strategies rather than a one-sizefits-all approach.To address this challenge, we focus on a fire-rescue scenario in which an operator must persuade a resident to evacuate as a high-stakes persuasion domain and propose Dialogue Policy Selection (DiPS), a Q-learning framework to dynamically select persuasion strategies adapted to the evolving conversational context.Specifically, we train a critic, trained to maximize the chance of evacuation success, to select a persuasion policy at each turn based on the resident's recent utterances.We then evaluate DiPS against multiple baselines in both simulated and real human interactions.We find that DiPS achieves higher evacuation success than a zero-shot LLM and generic RAG-augmented approach.* indicates equal contribution OP Hi.Do you need assistance leaving?RES I'm not leaving.I have too much work here.OP You need to leave.It's not safe.RES You really think it's that bad?OP It is.You must leave now.Need a vehicle
Institution: University of Southern California
Topic: Speech and dialogue systems
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7167212971
FWCI:
Cites:
Cited by:
Related to: