When it was 2024 – Generative AI in the field of digital scholarly editions

This contribution examines the current state of research on generative AI applications in digital scholarly editing. Drawing from experiments presented at the DHd 2024 workshop and additional literature, it identifies eight key application areas for Large Language Models (LLMs): (1) documentation of textual transmission, (2) post-processing of retro-digitized editions, (3) text establishment (transcription, layout analysis, OCR/HTR post-processing, markup), (4) normalization, (5) named entity recognition (NER) and other semantic annotation, (6) information enrichment, (7) translation, and (8) summarization. While among these areas, NER garnered the most experimental attention at the workshop, a comprehensive architecture for integrating generative AI across the full editorial stack was proposed. The paper concludes by identifying critical areas for future research: from a practical perspective, the field needs standardized workflow orchestration and evaluation protocols; from a theoretical perspective, researchers must systematically assess the strengths and weaknesses of LLMs in digital scholarly editions while addressing their inherent biases and ethical implications.