Answers are grounded in this paper and cite their sources.
Generative AI (GAI) can support writing, but learners may respond differently to its guidance. This study examined whether field dependence–independence (FDI) moderated the effect of GAI guidance on critical thinking behaviors (CTB) during human–AI collaborative writing. In a mixed-design study, 65 journalism students completed news-writing tasks with Standard GAI, which provided direct responses, and Heuristic GAI, which elicited learners’ reasoning. We coded 696 user prompts as CTB or non-CTB and analyzed them with logistic generalized linear mixed-effects models. Heuristic GAI was associated with more CTB overall, but its effect decreased as GEFT scores increased: students toward the field-dependent end showed larger increases in CTB and greater uptake of cognitive guidance. Source-specific analyses located this heterogeneity primarily in learners’ uptake of cognitive guidance; neither the condition nor moderation effect was significant for learner-initiated CTB. The findings support learner-sensitive GAI scaffolding in educational HCI.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: