Students with the best self-taught understanding of how language models work used AI tools most deliberately and achieved the highest assignment scores¶

Claim · All claims
Evidence · 1 study · causal r1 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Understanding of the language models, which was in every case self-taught, was associated with more effective prompts, more deliberate tool use, and the highest assignment scores. → Azimi 2026

Evidence¶

Azimi 2026¶

Azimi, S. (2026). AI Literacy over Tool Design: A Mixed-Methods Study of Scaffolded versus Unrestricted Generative AI in Programming Education. Delft University of Technology. https://arxiv.org/abs/2609.16784 doi:10.48550/arxiv.2609.16784

q2 · i? · causal · r1

Observational finding across the 33-student cohort: model understanding "varied widely across the cohort, and in every case it was self-taught, since the course did not teach it". The association with the highest scores is descriptive; no effect size is printed.

"The students with the best understanding of the models wrote more effective prompts and used the tools more deliberately, and they also achieved the highest assignment scores."

Discussion¶

Related Claims¶

• A scaffolded AI Study Coach produced no assignment-performance or concept-inventory advantage over unrestricted AI use in a master's programming course pilot — related
• Self-efficacy is positively associated with acceptance outcomes for AI-assisted English learning tools in adjusted regression models — related
• Students with a better understanding of how AI systems work show higher academic self-efficacy — related
• The Coach's hint budget was managed poorly: students split into early-spenders and savers, and within the Coach condition only those with a deliberate hint strategy scored higher — related