This study examines the integration of computational deep learning and digital literacy from 2011 to June 2025. Employing a hybrid methodology of Systematic Literature Network Analysis and Latent Dirichlet Allocation topic modeling, 141 high impact documents were synthesized following the PRISMA 2020 protocol. Findings reveal a conceptual shift from technical exploration (2011–2018) toward human-centric, pedagogical deep learning frameworks (2019–2025). While publications peaked in 2024, Australia and South Korea emerged as leading centers of excellence in citation impact. Latent Dirichlet Allocation modeling identified ten topics, uncovering a significant research gap in using AI for fundamental research processes compared to its dominance in instructional assessment. This study provides a novel mapping of thematic evolution and offers strategic recommendations for longitudinal empirical studies and inclusive AI-driven pedagogical designs.