The integration of digital technologies, artificial intelligence (AI), and data-driven systems has transformed Arabic language learning into a more accessible, adaptive, and interactive process. However, Arabic instruction in many educational contexts still relies on conventional teacher-centered approaches that limit learner engagement and autonomy. This study aims to identify major domains of technological innovation in Arabic language learning, analyze these innovations through a connectivist perspective, and propose a conceptual framework for AI-integrated instruction. Using a qualitative systematic literature review design, this study synthesizes recent scholarly publications related to educational technology, AI, gamification, adaptive learning, and connectivist pedagogy. The findings reveal four major domains of innovation: digital media integration, AI-based personalization, gamified learning strategies, and algorithm-driven instructional management. These innovations collectively support more learner-centered and network-based learning environments. Based on the findings, this study proposes the Connectivist AI Framework for Arabic Language Learning (CAF-AIL Model), which conceptualizes learning as a dynamic interaction among learners, digital platforms, technological systems, and algorithmic feedback mechanisms. Within this framework, technology functions not merely as a supporting tool but as an active node that mediates knowledge construction, learner engagement, and instructional adaptation. Despite its pedagogical potential, several challenges remain, including limited technological infrastructure, unequal internet access, low digital literacy, and insufficient teacher readiness. Therefore, successful implementation requires systematic teacher training, institutional technological support, and pedagogical redesign. This study contributes theoretically by extending connectivist learning theory into AI-integrated Arabic language education and practically by providing a framework for developing adaptive digital learning environments.