Digital divide 2.0: AI literacy and the widening stratification in marginalized communities
Abstract
The paper presents digital divides as a cycle rather than as separate problems. The previous research documented first digital divide that is concerned on unequal access to devices, electricity, broadband, software, and technical support. The second layer focuses on the skills required to use AI productively, including prompt guidelines, source verification, data interpretation, privacy protection, and the ability to understand a system’s limits. Artificial intelligence is transforming how knowledge is produced, assessed, and distributed in education, but access to AI benefits is uneven. This paper examines Digital Divide 2.0, the widening inequality that emerges when communities differ not only in connectivity and hardware but also in AI literacy, data sovereignty, language representation, critical evaluation, and institutional power. Using qualitative content analysis of the academic, policy, and institutional sources assembled in the working paper, the study organizes evidence across seven themes: infrastructural access; digital colonialism and data sovereignty; second-level skills and competency; critical and epistemic inequality; linguistic and cultural exclusion; policy and governance gaps; and intersectional marginalization. The synthesis shows that AI can reproduce educational stratification through several linked mechanisms. Under-resourced schools may lack reliable connectivity, devices, technical support, and trained teachers. Even when access exists, learners may remain excluded from higher-value uses of AI because they lack prompt design, verification, data literacy, or opportunities to participate in system design. This study also argues that AI literacy should be treated as a public educational capability rather than an individualistic skill. Equitable policy therefore requires infrastructure, localized and critical curricula, teacher development, community participation, indigenous data governance, and enforceable accountability mechanisms. Without these measures, AI literacy and education is likely to intensify existing inequalities rather than function as an educational equalizer.