Tag: learning-science

AI learning systems should treat feedback timing as a design variable: some errors need immediate correction, while other moments benefit from reflection, self-checking, or productive struggle first.

AI tutoring should sometimes make learning harder on purpose: spacing, retrieval, interleaving, delayed help, and fading support can improve durable learning when difficulty stays achievable.