With the rapid expansion of scientific data and the transformative potential of artificial intelligence (AI), Geosciences have increasingly adopted AI-based approaches for modeling, prediction, and data-driven discovery. This study presents a comprehensive bibliometric analysis of AI research in Geosciences between 2003 and 2023, drawing on 7,032 publications indexed in Scopus. Using bibliometric tools such as VOSviewer, BibExcel, and Biblioshiny, we analyzed publication trends, authorship patterns, collaboration networks, leading journals, and keyword co-occurrences. Results show an average annual growth rate of 14.71% and 16.12 citations per paper. China leads in publication volume (2,044 papers), while the United States demonstrates the highest citation impact (21.6 citations per paper). Neural networks, machine learning, and remote sensing emerged as dominant research themes. Keyword co-occurrence analysis revealed growing emphasis on deep learning, data mining, and environmental applications. Despite progress, significant gaps remain in hydrology, volcanology, and seismology, highlighting directions for future research. This analysis contributes to scientometrics by mapping a rapidly evolving interdisciplinary field and identifying research gaps, collaboration dynamics, and thematic shifts that will inform researchers, policymakers, and funding bodies.