Revising a text is an essential part of the writing process, and cultivating revision skills is crucial for developing advanced writing abilities. Although AI-generated feedback offers personalized and immediate support for revision, empirical studies suggest that many learners do not engage in revision after receiving it. This article addresses this issue by shifting the analytical focus from the quality of the feedback to the requirements of the revision process itself. Drawing on process-oriented writing research, the paper conceptualizes text revision as a sequence of interrelated sub-processes, deriving the specific cognitive, motivational and strategic demands that learners face when revising their own texts. Against this theoretical background, two AI-based feedback tools, Khan Academy Writing Coach and FelloFish , are analyzed to determine the extent to which their feedback practices align with these requirements. The analysis reveals three central tensions: (1) the timing of AI feedback versus learners’ need for critical distance from their own texts; (2) the risk of diminished learner agency and motivation when core revision processes are outsourced to AI; and (3) insufficient embedding of revision within meaningful writing tasks and communicative goals. The article argues that, without explicit alignment with revision processes, AI-based feedback may inadvertently hinder rather than support learners’ engagement with revision. Implications for the design of AI feedback tools, writing instruction and future empirical research are discussed.
Helm et al. (Thu,) studied this question.