GenAI literacy predicted immediate independent writing performance after AI removal for Visual Data Integration, Critical Thinking, and overall Composite Score¶

Claim · All claims
Evidence · 1 study · causal r2 · q2 · i1 small
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i1 In ordinal logistic regression, standardised GenAI literacy significantly predicted AI-removal Visual Data Integration (OR=1.14), Critical Thinking (OR=1.15), and Composite Score (OR=1.11), but not Insightfulness, Organisation and Coherence, or Linguistic Quality. → Yueqiao Jin 2025

Evidence¶

Yueqiao Jin 2025¶

Yueqiao Jin, Kaixun Yang, Roberto Martinez-Maldonado, Dragan Gašević and Lixiang Yan. (2025). The Agency Gap in AI-Supported Writing: How Reactive and Proactive Agent Designs Shape Multimodal Reasoning. https://arxiv.org/abs/2507.04398 doi:10.48550/arxiv.2507.04398

q2 · i1 · causal · r2

Ordinal logistic regression on 79 participants' AI-removal task scores, controlling for mini-VLAT domain knowledge, agent condition, and AI-support score. Printed effect sizes are small odds ratios (OR = 1.14, OR = 1.15, OR = 1.11 per one-SD literacy increase), which the article calls "modest", raising the odds of a higher score by 11–15%.

"GenAI literacy significantly predicted three dimensions of immediate independent performance after AI support was removed:Visual Data Integration,𝑏= 0.13,𝑆𝐸= 0.06,𝑧= 2.07,𝑝=.039,𝑂𝑅= 1.14;CriticalThinking,𝑏= 0.14,𝑆𝐸= 0.07,𝑧= 2.18,𝑝=.029,𝑂𝑅= 1.15; and the overallComposite Score,𝑏= 0.10,𝑆𝐸= 0.05,𝑧= 2.15,𝑝=.032,𝑂𝑅= 1.11."

Discussion¶

Learner Variables¶

• Academic and Critical Literacy — outcome: instruction changes it

Related Claims¶

• Performance during the AI-supported task consistently predicted performance on the subsequent AI-removal task across all writing dimensions — related
• AI tools can enhance short-term task performance while undermining durable learning (the learning-performance paradox) — related
• No significant GenAI-literacy-by-agent-condition interaction emerged, so the formal models did not support different literacy slopes across designs — related