GeoGPT: The One-stop AI Research Platform for Geoscientists

Session Chairs:
Grant Boquet, Michael Stephenson, James Ogg, Tao Wang, Junxuan Fan, Yong Ge

Discover how AI is transforming geoscience research with GeoGPT, a free, open-source AI assistant built on large language models (LLMs) and trained specifically on geoscience literature. GeoGPT offers flexibility in choosing foundational LLMs and features high-precision, real-time document parsing using retrieval-augmented generation (RAG) technology. Participants will explore GeoGPT’s public geoscience library and learn how to integrate their own data for customized research. Continuously evolving through collaboration with geoscientists, GeoGPT helps tackle real-world scientific challenges. This hands-on workshop offers a practical introduction to GeoGPT’s powerful tools, including paleontology classification, document reading and extraction, geological map recognition and generation, idea development, 3D Earth modeling, and automatic knowledge graph construction. The course is structured around three segments: an introduction to GeoGPT, application examples such as the evolution of life and the spatial-temporal distribution of igneous rocks, and GeoGPT-based solutions like petroleum geology knowledge systems and fossil taxonomy agents. Each segment includes time for discussion, with optional breakout sessions for participants to explore their own projects or research questions. Attendees are encouraged to bring their own data and ideas—just a laptop is needed. Come ready to learn, collaborate, and imagine the future of AI-enhanced geoscience.