NEP-MultiSent: A large-scale multilingual dataset on National Education Policy (NEP) 2020

This dataset comprises 1,840,710 records of educational content related to Artificial Intelligence (AI) in education, collected from diverse sources across all Indian states and union territories. The dataset is uniquely structured to support research on AI integration in educational systems, with particular emphasis on alignment with India's National Education Policy (NEP) 2020.Each record contains five attributes: a unique identifier (ID), language of content (Language), full text content (Content), publication timestamp (Date Published), and geographic location (Place\/Location). The temporal coverage spans a nine-year coverage during the period 2016-2025, capturing the critical period of NEP 2020 adoption and implementation in Indian education.The dataset enables multiple research applications including: (1) Natural Language Processing (NLP) analysis of educational content; (2) Geographic disparities in technology adoption in education; (3) Temporal trend analysis of AI integration in curricula; (4) Multilingual education content analysis; (5) Policy impact assessment of NEP 2020; and (6) Regional comparison studies across Indian administrative divisions.The large scale (1.84 million records) and comprehensive geographic coverage (all states\/UTs) make this dataset particularly valuable for training machine learning models, conducting longitudinal studies, and informing evidence-based educational policy decisions. The dataset supports research aligned with NEP 2020's focus areas including technology integration, multilingual education, and equitable access to quality education.