Higher post-task trust in an AI assistant predicts lower appropriate reliance during AI-assisted programming problem-solving¶

Claim · All claims
Evidence · 1 study (2 entries) · associational r1–r2 · q3 · i3 large
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? In a regression, trust was a strong negative predictor of appropriate reliance (b = -3.65, p < .001, R2 = .173). → Griffin Pitts 2026 q3 i3 Trust correlated negatively with appropriate reliance (r = -.42) and task accuracy (r = -.38). → Griffin Pitts 2026 (2)

Evidence¶

Griffin Pitts 2026¶

Griffin Pitts, Neha Rani, and Weedguet Mildort. (2026). Trust and Reliance on AI in Education: AI Literacy and Need for Cognition as Moderators. Preprint, accepted to AIED 2026. https://arxiv.org/abs/2604.01114 doi:10.48550/arxiv.2604.01114

q3 · i? · associational · r2

Ordinary least squares regression in the Results section predicting appropriate reliance from post-task trust among 432 undergraduates. The article reports "b=−3.65" with p < .001 and R2 = .173; no standardized effect size is printed, so impact is null.

"Trust was a strong negative predictor of appropriate reliance (b=−3.65,SE= 0.39,t(430) =−9.48,p<.001,R 2 =.173), indicating that students who reported higher trust tended to make fewer decisions appropriately in this task."

Griffin Pitts 2026 (2)¶

Griffin Pitts, Neha Rani, and Weedguet Mildort. (2026). Trust and Reliance on AI in Education: AI Literacy and Need for Cognition as Moderators. Preprint, accepted to AIED 2026. https://arxiv.org/abs/2604.01114

q3 · i3 · associational · r1

Pearson correlations reported in the Results section of the same 432-participant study: "r=−.42" for appropriate reliance and "r=−.38" for task accuracy, both p < .001. A correlation of .42 exceeds the .37 threshold for a large effect.

"Trust was negatively associated with both appropriate reliance (r=−.42, p<.001) and task accuracy (r=−.38,p<.001)."

Discussion¶

Related Claims¶

• AI literacy negatively moderates the trust–appropriate reliance relationship, with the trust slope becoming more negative as AI literacy increases — related
• The trust–appropriate reliance relationship is non-linear, with the decline attenuating at higher trust and no reversal within the observed data — a narrower finding that bears on this claim
• Need for cognition negatively moderates the trust–appropriate reliance relationship, with the trust slope becoming more negative as need for cognition increases — related
• Programming self-efficacy and programming literacy did not significantly moderate the trust–appropriate reliance relationship — related
• Reliance, trust, and dependency are distinct constructs requiring separate measurement from AI-literacy totals — related