General Intellectual Humility Is Malleable Through AI-Mediated Reflective Dialogue

Abstract

General intellectual humility (GIH) -- the recognition that one's beliefs may be fallible and revisable -- is associated with improved reasoning, learning, and social discourse, yet is widely regarded as a stable trait resistant to intervention. We test whether GIH can be elevated through a conversational intervention that combines staged cognitive scaffolding with personalized Socratic reflection. In a randomized controlled experiment (N=400), participants engaged in a structured, LLM-mediated dialogue that progressed from conceptual understanding of intellectual humility to applying, analyzing, evaluating, and generating novel, self-relevant scenarios that instantiate it. Relative to a time-matched control, the intervention produced a systematic increase in GIH, reduced rank-order stability, and tripled the rate of reliable individual improvement. Crucially, these effects persisted over a two-week follow-up without detectable decay. The effects generalized across political affiliation and did not depend on baseline personality profile. These findings challenge the prevailing pessimism regarding the malleability of GIH and suggest that scaffolded, Socratic reflection delivered through structured dialogue can produce durable changes in general intellectual humility.

Figures (1)

• Figure 1: (A) Mean GIH trajectories by condition. Mean GIH across pre-intervention (baseline), immediate post-intervention, and 14-day follow-up, shown for treatment and control groups with standard errors of the mean. All available observations at each time point were used. Groups were comparable at baseline, diverged immediately after the intervention, and remained separated at follow-up. Dashed vertical segments mark the difference-in-differences (DiD) contrasts at post and follow-up. (B) Mixed-model treatment effects. Mixed-effects estimates of the treatment effect on GIH at immediate post-intervention and 14-day follow-up, shown as condition-by-time interaction coefficients. Both estimates are positive and statistically significant, and their near equivalence is consistent with persistence of the treatment effect without detectable decay. (C)-(D) Distribution of individual changes in GIH by condition. Left: immediate post-intervention change relative to baseline (post - pre). Right: 14-day follow-up change relative to baseline (follow-up - pre). Dashed vertical lines indicate group means. Across both cases, the treatment distribution is shifted rightward relative to the control distribution, indicating larger gains in GIH under the intervention.