<![CDATA[Generative AI tools capable of producing fluent text, synthesizing information, and delivering formative feedback have entered educational settings with a speed that has outpaced institutional policy, pedagogical adaptation, and scholarly understanding. Reshaping the terrain of writing instruction, assessment design, and academic integrity in ways that are simultaneously promising and deeply unsettling, these tools present educators and institutional leaders with a set of challenges that neither blanket prohibition nor uncritical adoption is equipped to resolve. This evidence-informed conceptual paper synthesizes scholarship on writing-to-learn, feedback for revision, academic integrity, and responsible AI governance to propose a practical, integrated framework for the responsible use of generative AI in writing instruction and feedback contexts. Drawing on research traditions in formative assessment, learning-oriented feedback, integrity by design, and responsible AI principles, the paper articulates four interdependent domains: (a) pedagogical use cases grounded in clearly specified learning goals; (b) transparency and disclosure norms supported by systematic AI literacy development; (c) assessment redesign oriented toward process evidence and authentic performance; and (d) governance, privacy, and equity safeguards embedded in institutional procurement and policy frameworks. Three conceptual tables operationalize the framework by providing a use-case taxonomy with associated risks and guardrails, an assessment redesign menu calibrated for integrity and learning in AI-present contexts, and a policy checklist for institutions and journals. The paper concludes with targeted recommendations for teachers, institutional leaders, and quality assurance bodies seeking to harness generative AI as a genuine learning resource while protecting student agency, privacy, and the social trust upon which credentialing ultimately depends.]]>
