Exploring thematic structures and emerging trends in open access AI research in India using LDA

The field of artificial intelligence (AI) is one of the rapidly growing areas of research, and open access (OA) publishing is a key factor in promoting knowledge sharing and dissemination. This study examines the growth, citation impact, and thematic development of open-access artificial intelligence (AI) research in India using a combined bibliometric and text-mining approach. The Scopus database was searched to retrieve data on OA publications that were based in India. A total of 18,625 publications spanning 1986 to 2025 were retrieved, accumulating 362,236 total citations (mean: 19.45 citations per paper; median: 6.0). The publication and citation trends were analysed using bibliometric methods, whereas Latent Dirichlet Allocation (LDA) was applied to identify key themes in the research. The findings indicate a significant increase in publication output, particularly from approximately 2015 onwards, with the most remarkable surge occurring after 2018—annual output jumped from 169 publications in 2017 to 414 in 2018 (+ 145%), and reached a peak of 4604 publications in 2024. Six major thematic clusters were identified, with deep learning and neural networks emerging as the dominant theme by publication volume, while AI in healthcare demonstrates the highest average citation impact per paper. Within this OA corpus, citation impact varied considerably across themes, reflecting differences in interdisciplinarity, field maturity, and application orientation. The study does not include a non-OA comparison group and therefore does not estimate a causal OA citation advantage; observed patterns are consistent with multiple contributing factors including topic prominence and citation time-lag. Overall, the study provides insights into the thematic evolution and research impact of open-access AI research in India.