A new blog post by Felix Vanden Borre for the VOICES Project examines how AI can support historical research through the use of Named Entity Recognition (NER), while also highlighting the challenges posed by complex historical texts.
The VOICES Project is exploring a range of computational methods to analyse historical sources at scale and uncover the lives of non-elite women in early modern Ireland. One of the key approaches being tested is Natural Language Processing (NLP), a branch of artificial intelligence focused on enabling computers to understand and analyse human language.
Among the most useful NLP techniques for historians is Named Entity Recognition. NER systems are designed to identify and classify references in texts, including people, places, political or religious groups, currencies, dates and works of art. By automatically detecting these entities across large collections of documents, researchers can rapidly identify patterns, connections and areas of interest that might otherwise remain hidden.
However, the blog notes that applying NER to historical sources is not without challenges. Historical texts often contain archaic language, inconsistent spelling and complex references that can make automated analysis difficult. While NER offers significant gains in efficiency, it can also miss important entities or misclassify information.
Despite these limitations, Named Entity Recognition remains a powerful tool for digital humanities research. Modern NER models are generally considered state of the art when they achieve accuracy rates of 90 per cent or higher, demonstrating both the potential and the ongoing challenges of using AI to investigate the past.
The full blog post explores how the VOICES Project is testing these technologies to better understand historical records and recover previously overlooked stories from Ireland’s early modern period. Read it here.
