["TESOL Quarterly, EarlyView. ", "\nAbstract\nThis Research Issues paper outlines the theoretical and practical foundations for assessing GenAI‐mediated DMC, with a particular focus on frameworks for ethical authorship, diverse assessment approaches and methods, and flexible quality criteria and rubrics. Drawing on theoretical traditions including social semiotics, systemic functional linguistics, multiliteracies, and sociocultural theories, we have emphasized the importance of ethical and responsible authorship in AI‐mediated composition. In particular, we argue that fine‐tuned prompt literacy and transpositional grammar allow learners to interact critically and creatively with GenAI tools, enabling them to navigate complex semiotic environments while maintaining agency and ethical authorship. Successful assessment in this context demands flexible, context‐sensitive models that reflect the evolving nature of GenAI‐mediated DMC. Rather than relying on static rubrics or prescriptive quality metrics, we advocate for co‐constructed assessment criteria grounded in ethical engagement, iterative learning, and critical reflection. We note in the implications that more scholarly attention is needed to explore what assessment strategies for ethical authorship in AI‐mediated DMC look like in practice, and how teachers can foster critical thinking and reflection on the metalanguage of multimodality beyond simply understanding the affordances of modes and media to achieve specific communicative functions.\n"]