Editors’ Summary: This post details a workflow to use AI to assist with archival research. The author, a PhD candidate in history, discusses returning from a trip to the archives with hundreds of cell phone images of documents and struggling to organize them. He describes his method of using Zotero and Claude to organize these images into text searchable PDFs with structured metadata. The author acknowledges that LLMs are not a good substitute for the historian’s essential functions of close reading, critical analysis, and writing, but can be useful for extracting structured data from images.
Editors’ Choice: AI in the Archive: Saving Time by Turning Photos into Structured Metadata
