This systematic literature review examines the feasibility of AI-based screening tools as alternatives to South Africa's Policy on Screening, Identification, Assessment and Support (SIAS, 2014) for identifying neurodivergent learners in under-resourced mainstream schools. The review is guided by the neurodiversity paradigm and Universal Design for Learning (UDL). Searches will be conducted across six databases: Scopus, PubMed, PsycINFO, African Journals Online (AJOL), Sabinet African Journals, and Africa-Wide Information (NISC), covering peer-reviewed publications from 2014 to 2026. Findings will be synthesised thematically across three strands: SIAS implementation challenges, AI screening tool capabilities, and feasibility of AI tools in low-resource South African school contexts. Reporting will follow the PRISMA 2020 guidelines.