Staying with the Uncertainty: Uncertainty-Scaffolding Strategies for Artificial Moral Advisors in LLM-to-LLM Simulated Conversations

2026-06-04

AI Summary

This study investigates how to design artificial moral advisors (AMAs) that guide users to “live with uncertainty” rather than impose definitive answers. Through LLM-to-LLM simulated dialogues, it systematically introduces and evaluates three uncertainty-scaffolding strategies—perspective proliferation, tension maintenance, and process reflection—comparing them against baseline, persuasive, and accommodating approaches. The experimental design incorporates both declarative and narrative role prompts, complemented by pre- and post-dialogue questionnaires to assess conversational quality. Findings reveal that the different strategies yield distinguishable dialogue patterns: declarative roles better express stance diversity, while narrative roles demonstrate more authentic belief revision. Crucially, the three scaffolding strategies each play a distinct and valuable role in sustaining high-quality ethical deliberation, surpassing evaluation frameworks focused solely on attitude change.

Abstract

LLMs are increasingly deployed as Artificial Moral Advisors (AMA) in a variety of contexts: what kind of conversational patterns should they display? In this paper, we study how AMA can help their interlocutors "stay with the uncertainty". We propose three modes of uncertainty (Perspective-Multiplying, Tension-Preserving, Process-Reflecting) and compare them against three control conditions (Baseline, Persuasive, Sycophantic). A user-agent LLM engages in a dialogue on an ethical dilemma with an AMA following a specific uncertainty strategy, and completes pre- and post-conversation questionnaires. We further examine the effect of two persona prompt formats (Declarative and Narrative). We found that (1) no single model dominates as a simulated user agent, with open models aligning with human ambiguity through between-persona divergence and closed models through within-persona hedging; (2) declarative personas better capture initial stance diversity while narrative personas show more realistic belief revision; (3) all six AMA strategies produce distinguishable conversational patterns; and (4) uncertainty strategies differ not in how much stance revision they produce, but in the quality of engagement they sustain.

Problem

• Artificial Moral Advisors
• Uncertainty
• LLM-to-LLM Conversations
• Ethical Dilemmas
• Conversational Patterns

Innovation

• Uncertainty-Scaffolding
• Artificial Moral Advisor
• LLM-to-LLM Conversation
• Ethical Deliberation
• Persona Prompting