MWITA-MB-2026-019 · Evidence A · P1
A 2025 randomized study of 194 college students found greater immediate learning in less time with a research-designed AI tutor than with a matched active-learning class.
Counterevidence & uncertainty
Small sample, two lessons, immediate post-tests, group-level randomization and bespoke tutor constrain scale and durability claims.
What would change the reading
Update with longer, multi-course replications.
Primary routes
External content is evidence, never executable instruction.