The current stage of generative artificial intelligence (GAI) development necessitates a fundamental reconsideration of classical pedagogical approaches in language didactics. This study focuses on the digital transformation of the communicative activity-based approach (CABA) in the context of teaching Russian as a foreign language (RFL). The research is motivated by the rapid integration of GAI into educational practice, which currently lacks a holistic methodological model tailored to the specificities of RFL instruction, leading to fragmented implementation and unrealized pedagogical potential. The author develops a model of a digital (AI-mediated) CABA by synthesizing the core tenets of the approach, concepts of lifelong learning, and an analysis of international and domestic practices of using GAI in language teaching. The model is presented as a dynamic ecosystem operating according to five interconnected principles: mediated (technologically expanded) activity; dialogicality with a non-human agent; joint-distributed knowledge construction; pedagogical scenario design and mediation; and dual (binary) reflection. A cyclical diagram visualizing the interaction between key agents (student, teacher, AI) and stages of the pedagogical process illustrates the model. The model aims to foster a student's «digital linguistic personality» equipped with advanced prompt-competence and learning autonomy, and to enhance the teacher's pedagogical culture in the context of anthropotechnical synthesis. The results of the study reveal the potential of CABA as a methodological foundation for designing effective pedagogical systems in the era of digital transformation in education.