Learners rated the L2C coach higher than both baselines on safety, performance, agency, and satisfaction¶

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

Subclaims¶

q3 i? Post-trial questionnaire ratings for safety and performance were higher for L2C than both baselines, consistent with objective improvements. → Wei Wang 2026

Evidence¶

Wei Wang 2026¶

Wei Wang, Enlin Gu, Antonio Loquercio, Haimin Hu, Rahul Mangharam. (2026). AI Coaching for Accelerating Human Skill Development with Reinforcement Learning. arXiv preprint. https://arxiv.org/abs/2606.25337 doi:10.48550/arxiv.2606.25337

q3 · i? · causal · r2

Post-trial 5-point Likert questionnaire (Table 2 items) collected after Phase 3 of the user study. The article reports L2C "received higher subjective rat- ings than both baselines onsafety" and performance, and adds that higher agency and satisfaction ratings suggest adaptive coaching preserves user control.

"On the post-trial questionnaire, L2C also received higher subjective rat- ings than both baselines onsafety(“Compared to my first race, I flew more safely in the final race.”)"

Discussion¶

Related Claims¶

• L2C outperformed both the Rule-based Fading and Minimally-invasive Assistance baselines on lap time and failure count — related
• L2C coaching significantly reduces learners' lap times and failure counts from pre- to post-coaching in FPV drone racing — related
• Baseline coaches produced weaker or absent learning gains: RBF improved only failure count, MIA neither outcome — related