Learning outcomes differ by how students use AI: augmentation and mixed users show gains in both sessions, while automation users' Session One essay advantage is absent from their unaided essays¶

Claim · All claims
Evidence · 1 study · causal r3 · q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? Of treated AI users, 61 percent are classified as augmentation users, 9 percent as automation users, and 27 percent as mixed users; augmentation and mixed users show positive test-score differences relative to control in both sessions, and automation users' higher-rated Session One essays do not carry over to their unaided Session Two essays. → Zara Contractor and Germán Reyes 2026

Evidence¶

Zara Contractor and Germán Reyes 2026¶

Zara Contractor and Germán Reyes. (2026). Experimental Evidence on the Learning Impact of Generative AI. https://arxiv.org/abs/2607.08849 doi:10.48550/arxiv.2607.08849

q3 · i? · causal · r3

Descriptive between-group comparisons from the randomized experiment, classifying treated AI users' ChatGPT conversation logs as automation, augmentation, mixed, or other. The article reports these patterns are "consistent with the model’s prediction that what students learn from AI depends on how they use it."

"Automation and mixed users submit higher-rated essays in Session One, but automation users’ advantage is absent from their unaided essays one week later. Augmentation and mixed users show positive test-score differences relative to the control group in both sessions"

Discussion¶

Related Claims¶

• How genAI is used, not who uses it, predicts whether independent performance declines or improves — the general claim this finding bears on, with the studies behind it
• AI access raises unaided essay quality in both sessions, while AI-generated text detectable in Session One essays disappears by Session Two — related
• One week later, the AI-access group's lead on unaided assessments shrinks to 3.9 pp (0.21 SD), short of statistical significance (p = 0.068) — related
• As cited second-hand in an AI-literacy framework paper: ChatGPT use improved short-term essay scores but triggered metacognitive laziness with no significant knowledge-transfer gains — related
• Product-outcome studies report GenAI-related gains on product-quality measures but do not assess durable, transferable learning — related
• Students expect AI to help their test scores, but control students overestimate the effect relative to the experimental estimate — related