AI in the Humanities & Cultural Heritage
AI is reshaping how scholars work with the historical record. This hub explores what these tools can and can't responsibly do in the humanities, from opening closed archives to the risks of trusting a machine with the past.
8 articles
Leo Team · August 18, 2026
AI for Historical Ledgers: What Works, What Breaks, and How to Build a Defensible Pipeline
How to build a defensible structure-first AI pipeline for handwritten historical ledgers, where cell assignment failures matter more than character error and no transferable benchmark yet exists.
Read the articleLeo Team · August 18, 2026
Historical Text Translation AI: What It Does Well, Where It Fails, and How to Use It Defensibly
What historical text translation AI does well and where it fails, why transcription, normalisation and translation must stay separate, and how researchers can verify output defensibly.
Leo Team · August 18, 2026
Historical Transcription to CSV and JSON: Turning Transcribed Records into Data You Can Analyse
How to turn historical transcriptions into CSV or JSON by keeping the faithful text as master and deriving analytical tables so every value stays recoverable.
Leo Team · July 29, 2026
Historical Entity Extraction: A Practical Guide for Historians Who Answer to Peer Review
A practical guide to historical entity extraction (NER/NEL) covering vocabulary, method choice by era/language, the outsized impact of transcription quality on downstream accuracy, authority-file selection, and TEI encoding for a defensible pipeline.
Leo Team · July 29, 2026
Prosopography AI: Turning Transcribed Records into Person Data
Explains the five-stage pipeline (HTR, NER, coreference, record linkage, structured modelling) for building prosopographical person data from archival records, arguing transcription quality bounds accuracy downstream and warning that LLMs fabricate plausible but nonexistent people if unverified.
Leo Team · July 26, 2026
AI-Generated Finding Aids: What the Technology Can Actually Do, and Where It Still Needs You
Breaks down "AI-generated finding aids" into six distinct sub-tasks (HTR/OCR, NER, scope-note drafting, arrangement, authority reconciliation) and assesses which are production-ready versus pilot-stage, arguing transcription quality is the foundation everything else inherits.
Leo Team · July 20, 2026
AI and the Humanities: How Computation Supports Historical Scholarship Without Replacing It
AI in the humanities as a controlled scholarly tool for transcription, pattern-finding, and translation, with historians retaining verification, interpretation, and source criticism.
Leo Team · July 20, 2026
AI in Historical Research: What's Reliable and What Isn't
Historical research and AI: which tools reliably transcribe, translate, and extract names from sources, and which generative systems risk plausible fabrications in readings and citations.
Decipher your archive
© 2026 Leo Technologies Limited. All rights reserved