Algorithmic Mentors and Multilingual Writing Development

This chapter reinterprets generative AI within the framework of Activity Theory (AT), transitioning from an independent text creator to a designated 'algorithmic mentor' that supports diagnostic reasoning, genre-sensitive explanation, and metacognitive uptake in multilingual revision. It shares a theory-building, design-based account to update writing studies that link sociocultural writing development, peer-mediated revision, translanguaging and growing literature on feedback in AI-supported environments. It suggests design principles for eight projects to situate a five-phase “trialogue” coordinating authorial intent, peer diagnosis, and AI consultation. K -12/university scenarios show equity-oriented orchestration, transparency protocols, and process-focused assessment approaches. Contributions involve an AT model of AI-mediated writing practice, a routine that privileges explanation over replacement, multilingual designs validating translanguaging/code meshing in writing for multilingual students, and governance principles for privacy, fairness, and integrity.