Using AI to Read Historic Letters
This exhibit demonstrates how artificial intelligence can be used to transcribe historic handwritten letters through Handwritten Text Recognition (HTR). By revealing the model’s reasoning process, it shows how closely AI interpretation mirrors the trained eye of a paleographer.
Unlocking AMC Early Presidents
How does a machine learn to read the past? This digital exhibit explores the intersection of archival history and Artificial Intelligence through the personal correspondence of Texas A&M’s early presidents. By utilizing Handwritten Text Recognition (HTR) and the IIIF Paleography utility, we do more than just transcribe "ink on page." We leverage Google Gemini to reveal the AI’s "Chain of Process" reasoning—demonstrating how a modern model interprets 19th-century cursive, navigates historical shorthand, and mirrors the trained eye of a professional paleographer. Step inside to see the letters, the transcriptions, and the digital logic that brings them to life.


