["British Journal of Educational Technology, EarlyView. ", "\nAbstract\n\nWhile generative artificial intelligence (GenAI) tools demonstrate potential for enhancing students' learning outcomes, little is known about how students use GenAI at the micro‐level, particularly regarding when and how they seek assistance during self‐regulated learning (SRL). This study examined how learners' achievement goal orientations influence GenAI use patterns during problem‐solving within an AI‐powered learning environment. A total of 114 university students completed a nutrition recommendation task on the Healthy Choice platform. Prior to the task, students' goal orientations were measured using a self‐report scale. During the task, system log files captured learners' SRL activities and GenAI interactions. Hierarchical clustering analysis identified three distinct learner profiles: mastery‐oriented, performance‐oriented and low‐goal learners. Mastery‐oriented and performance‐oriented learners outperformed low‐goal learners. While ANOVA results revealed no significant differences in GenAI usage frequency across clusters, epistemic network analysis demonstrated significant differences in how GenAI was integrated into SRL processes. Mastery‐oriented learners exhibited stronger connections between GenAI use and cognitive activities (execution and evaluation), leveraging AI to deepen conceptual understanding. Performance‐oriented learners primarily used GenAI to support initial decision‐making. In contrast, low‐goal learners showed stronger temporal associations between GenAI use and metacognitive tasks like monitoring, reflection and final decision‐making. These findings inform differentiated scaffolding approaches based on student motivation profiles and the design of adaptive learning technologies that support personalized, effective engagement with GenAI tools.\n\n\n\n\nPractitioner notes\nWhat is already known about this topic\n\nGenerative artificial intelligence (GenAI) tools can generate adaptive educational content and feedback.\nPrevious studies have explored the positive impacts of GenAI integration on student learning experiences and outcomes.\nLearners' achievement goal orientation (AGO) has the potential to influence their GenAI use patterns.\n\nWhat this paper adds\n\nThis study revealed how students with varying goal orientation profiles differed in the temporal patterns of GenAI use during their self‐regulated learning (SRL) processes.\nMastery‐oriented learners tended to request assistance from GenAI tools as they conducted SRL activities of execution and evaluation, while performance‐oriented learners utilized GenAI tools for initial decision‐making.\nLow‐motivated learners relied on GenAI tools for metacognitive activities of monitoring and reflection, potentially leading to cognitive dependency.\n\nImplications for practice\n\nFrequency of GenAI use matters less than when and how students use these tools.\nDifferent scaffolding approaches are needed based on student motivation profiles.\nUnderstanding GenAI use patterns helps design AI‐supported learning environments that match student needs and promote meaningful learning rather than cognitive offloading.\n\n\n\n\n"]