Digital Education and E-Learning Innovations

Background: Generative artificial intelligence is becoming a natural part of everyday digital platforms, including WhatsApp, Google Workspace, and Canva. AI tools are now available even in low-resource environments. However, there is very little information on the actual utilisation of these technologies in Indian Non-Governmental Organisations (NGOs), especially among student beneficiaries. Objective: This study aimed to analyse access, understanding, and utilisation of generative AI tools by three audiences, including student beneficiaries of NGOs, student teachers and staff of NGOs, and student learners in digitally advantaged mainstream schools. Method: A cross-sectional descriptive study was carried out across various NGOs in India including Calcutta Foundation, Kshamata, Children Hope India (Delhi and Hyderabad), Akaanksha Foundation, Kailash Satyarthi Foundation, Sahaara, Kurniv Foundation, and Chaitanya Kul, with a comparative sample of mainstream schools. The sample size was 157 (NGO student beneficiaries: n=111, NGO teachers/staff: n=9, mainstream school students: n=37) selected through convenience sampling. Descriptive statistics were used to analyse data, including frequencies and percentages, and comparative analysis on a group level. Results: Findings showed that there was a high level of inequality in the access to AI in the NGO context, with 9% of NGO student beneficiaries reporting that they were not allowed to use their phones at all because of institutional policies. The academic purpose was the most prevalent, and more than 55 NGO students cited studies as their main purpose of using AI tools. In general, 78.1% of NGO beneficiaries said that AI had a positive influence on their studies. The most commonly used tools by NGO students were Meta AI on WhatsApp and ChatGPT, whereas mainstream school students showed a wider range of multi-platform applications and were more critical. The most common challenges that were mentioned included poor internet connectivity (reported by 25 NGO students) and language barriers. Conclusion: This study highlights that while embedded GenAI tools lower entry barriers, unequal institutional access, infrastructural limitations, and limited digital literacy support may reinforce educational inequalities rather than reduce them.