GraphRAG and the Ancient World: Toward a Spatial AI Research Platform

Digital Classicist London seminar 2026 Institute of Classical Studies, University of London, Senate House WC1E 7HU. Fri, 26 Jun 2026, Federico Di Pasqua (Zurich), GraphRAG and the Ancient World: Toward a Spatial AI Research Platform. This paper presents a spatial AI research platform for the Ancient World built around a GraphRAG architecture designed to address a central problem in digital classics, ancient history, and archaeology: the fragmentation of evidence across heterogeneous and only partially interoperable resources. Literary texts, inscriptions, papyri, archaeological datasets, gazetteers, and modern scholarship are increasingly available in digital form, yet they remain difficult to query in ways that preserve provenance, context, and the relations among texts, places, entities, and interpretations. I argue that GraphRAG provides a rigorous and scalable framework for this scholarly environment. Rather than relying on vector retrieval alone, the platform integrates large language models with a knowledge graph encoding structured relations among passages, works, places, periods, source types, and domain concepts. A key feature of the system is its integration with gazetteers and Linked Open Data infrastructures, which supply stable identifiers, spatial authority, and interoperable links across otherwise dispersed corpora. In this way, retrieval becomes not only semantic, but also spatial and relational. The result is a research environment in which users can move between map-based exploration, graph-based discovery, and natural-language querying, while remaining anchored to traceable source passages and explicit retrieval paths. The aim is not simply to generate fluent responses, but to build AI-assisted infrastructure that remains auditable, citation-grounded, and philologically controlled. More broadly, the paper proposes GraphRAG as a mature model for integrating ancient-world datasets within a transparent and reusable spatial AI environment, and as a methodological contribution to current debates on how large language models can be adapted for rigorous humanities research. 🎙️ New to streaming or looking to level up? Check out StreamYard and get $10 discount! 😍 https://streamyard.com/pal/d/60799986...