AI Interaction Competence, not GPA or prior knowledge, is the decisive moderator of LLM-assisted learning gains¶

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

Subclaims¶

q3 i? The Treatment × AIC interaction was positive and significant, with high-AIC participants benefiting disproportionately. → Bharat Anand 2026 q3 i? In the three-way interaction, novices with high AIC benefited more from LLM-assisted study, while novices with low AIC showed substantially smaller gains. → Bharat Anand 2026 (2)

Evidence¶

Bharat Anand 2026¶

Bharat Anand, Lihi Idan. (2026). Generative AI and the Productivity Divide: Human–AI Complementarities in Education and Knowledge Work (Working Paper). https://arxiv.org/abs/2605.18143 doi:10.48550/arxiv.2605.18143

q3 · i? · causal · r2

Interaction analysis in the randomized experiment; AIC was inferred from behavioral performance rather than self-assessment. The article states AIC supplanted traditional academic credentials as the most consequential moderator of performance.

"In interaction analysis, the coefficient on the (Treatment×AIC) term was positive and significant (p<0.05)."

Bharat Anand 2026 (2)¶

Bharat Anand, Lihi Idan. (2026). Generative AI and the Productivity Divide: Human–AI Complementarities in Education and Knowledge Work (Working Paper). https://arxiv.org/abs/2605.18143

q3 · i? · causal · r2

Three-way interaction analysis including Treatment, pre-intervention exam score, and AIC; after rescaling the pre-intervention score the coefficient was positive, indicating novice learners benefited more from LLM-assisted study only if they had high AIC levels.

"The interaction term was statistically significant (p<0.05)."

Discussion¶

Related Claims¶

• GPA and prior topic knowledge do not moderate the performance effect of LLM-assisted self-study — related
• Preference for LLMs aligns only weakly with AI Interaction Competence, producing enthusiastic-but-unskilled and skilled-but-reluctant subgroups — related
• LLM-assisted self-study yields higher post-intervention exam performance than traditional resources among early-career knowledge workers — related
• Baseline capability moderates AI productivity benefits non-monotonically across AI's capability frontier — a broader claim this one bears on
• Access to a generative-AI assistant raises task performance for both lower- and higher-education adults, with significantly larger gains for lower-education participants — related
• In one randomized working-paper experiment, conceptual-map scaffolding gave LLM-using novices marginally higher scores (p < .10) and lower score variance, with a marginal benefit concentrated among low-AIC novices — reports the opposite
• Extending mandated daily study time from three to four hours produces no significant performance gains for novices — related
• Exploratory, non-significant pattern suggests students with stronger learner profiles may show larger AI-assisted exam gains — related
• Accounting for learning behavior attenuates profile-related exam score differences, eliminating significance in Python — related
• Learning gains were uncorrelated with prior self-reported AI literacy and database literacy, suggesting AI integration benefited students irrespective of prior perceptions — reports the opposite
• Prior physics knowledge moderated the correlation between feeling of learning and actual test performance (positive interaction, P < 0.05) — related
• Self-questioning during study benefits low verbal ability high school students more than high verbal ability students (Experiment 1) — related
• Denial of LLM access increases attrition and LLMs are the most preferred study resource, acting as revealed preference — related
• No reliable moderation of depth effects by baseline knowledge, though simple slopes suggest an exploratory pattern — related
• Teachers' initial SRL competence profiles moderate training effectiveness: initially high-competence teachers benefit more from training — related
• Not all students would benefit from improving the same competencies, suggesting an individualized approach to education may be preferable to a one-size-fits-all approach — a broader claim this one bears on