Confidence Amplification in Human–AI Moral Decision-Making

Abstract

As artificial intelligence (AI) systems increasingly engage users in value-laden discussions, a key concern is whether they shape not only what people decide but how confident they feel. We examined confidence amplification, shifts in certainty without decision reversals, in human–AI moral decision-making. In a 2 (Time: pre vs. post) × 3 (AI moral framing: utilitarian, deontological, balanced) mixed design, 120 participants engaged a 10-round conversation with an AI assistant discussion an autonomous-vehicle dilemma. Decision reversals were uncommon and unaffected by framing. However, confidence increased reliably from pre- to post-discussion, and a confidence-weighted opinion-strength measure shifted toward the AI’s framing. Echo-chamber dynamics emerged when framing aligned with participants’ initial stance: aligned participants gained confidence whereas misaligned participants did not, with parallel increases in linguistic certainty. These findings suggest that amplification of confidence, rather than directional opinion change, may be a central mechanism through which AI influences moral judgment.