Using the AIED-U system yielded a positive indirect effect on learning gains through a reduction in the effort teachers reported¶

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? Robust mediation analysis showed a statistically significant indirect effect of the experimental (vs training) condition on learning gains through the Effort subscale (AB = 3.649, 95% CI [1.097, 6.398], p-adj = 0.007). → Luiz Rodrigues 2026

Evidence¶

Luiz Rodrigues 2026¶

Luiz Rodrigues, Guilherme Guerino, Thomaz Edson Silva, Luana Bianchini, Mariana Alves, Geiser Chalco Challco, Thales Vieira, Marcelo Marinho, Valmir Macario, Diego Dermeval, Ig Ibert Bittencourt, Seiji Isotani. (2026). AIED unplugged, teacher workload, and numeracy learning: a clustered quasi-experimental mixed-methods study. Smart Learning Environments. https://doi.org/10.1186/s40561-026-00465-x

q3 · i? · causal · r2

Robust mediation analysis tested whether the experimental condition (relative to training) affected each workload subscale, which in turn affected learning gains; the Effort path yielded AB = 3.649, CI [1.097, 6.398], p-adj = 0.007, the only subscale significant after adjustment.

"The results suggest a statistically significant mediation effect for Mental Demand and Effort, although only the latter remains significant after p-values adjustment. Overall, this finding suggests that the experimental condition reduces teachers' effort, which in turn is negatively associated with students' learning gains."

Discussion¶

Related Claims¶

• Teacher perceived workload did not differ significantly across conditions on any NASA-TLX subscale, though the experimental condition was descriptively lower on most subscales — related
• Students taught by teachers who received training outperformed those in the control condition, and adding the AIED-U system did not improve learning gains further beyond training alone — related
• Teacher perceived workload was negatively associated with student learning gains, especially for Mental Demand and Effort — related
• AI-supported co-produced feedback achieved comparable quality and statistically significant learning gains with reduced production time and workload — related
• The AI+Robot condition is associated with lower perceived workload than AI-Only on effort, frustration, temporal demand, physical demand, and Raw TLX — related
• With the alerting & guiding dashboard, teachers' workload dropped most on the performance and frustration subscales, while cognitive load, temporal demand and effort differences were non-significant — related