The majority of AI-assisted participants self-report using the AI primarily to get answers directly, a pattern validated against chatlogs¶

Claim · All claims
Evidence · 1 study · causal r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? 61% of AI-condition participants in Experiment 2 reported using the AI primarily to get answers directly; 27% for hints or clarifications; 12% reported no usage. → Grace Liu 2026

Evidence¶

Grace Liu 2026¶

Grace Liu, Brian Christian, Tsvetomira Dumbalska, Michiel A. Bakker, Rachit Dubey. (2026). AI Assistance Reduces Persistence and Hurts Independent Performance. Published as a conference paper at COLM 2026. https://arxiv.org/abs/2604.04721 doi:10.48550/arxiv.2604.04721

q2 · i? · causal · r2

Self-report multiple-choice question at the end of Experiment 2's AI condition. Chatlog records validated the reports: of 37 self-reported non-users, 83.8% did not message the AI at all and 94.1% messaged fewer than three times.

"We found that the majority of participants (61%, N= 189) in the AI condition self-report that they used the AI primarily to get answers directly."

Discussion¶

Related Claims¶

• Persistence and performance costs concentrate among participants who used the AI to obtain direct solutions, while hint-users resemble controls — related
• AI assistance reduces persistence: persistence costs concentrate among learners who used AI for direct solutions, not hints — related
• Students use the integrated AI assistant differently across writing stages: clarification early, verification late — related
• Engineering students most often use LLMs for academic writing (27%), conceptual clarification (20%), programming help (17%) and brainstorming (13%) — related