Integrating GenAI feedback and dialogic peer feedback to improve academic writing performance and revision practices: a quasi-experimental study

While GenAI offers unprecedented efficiency in delivering feedback, its use often reinforces transmission-oriented models that position undergraduate students as passive recipients. This study investigated how combining GenAI feedback with dialogic peer feedback influenced undergraduate students’ academic writing performance and revision behaviours through a quasi-experimental design. The results revealed that the combination of GenAI feedback and dialogic peer feedback most effectively enhanced academic writing performance and produced more sustained improvement. In addition, the GenAI feedback combined with dialogic peer feedback generated a higher frequency of full and extra revisions, whereas reliance on GenAI feedback alone more often resulted in partial revisions or non-implementation. Semi-structured interviews revealed distinct responses across conditions: GenAI feedback alone provided initial structural guidance but imposed cognitive burden and limited interpretive support; in contrast, GenAI feedback combined with dialogic peer feedback created a shared interpretive space that enabled collaborative negotiation and supported undergraduate students’ transformation from passive recipients into agentic participants. These findings advocate for student-centered approaches to GenAI integration, demonstrating how collaborative dialogue transforms GenAI output from prescriptive information into meaningful learning resources.