Digital Humanities 2023 More than Meets the (Artificial) Eye: Exploring Historical Photographs from Ireland with Computer Vision Methods Osti, Giulia giulia.osti@ucdconnect.ie University College Dublin Cushing, Amber amber.cushing@ucd.ie University College Dublin Little, Suzanne suzanne.little@dcu.ie Dublin City University Historical photographs are value-laden artefacts. Viewed as a "certificate of presence" (Barthes 1993: 87) of people, places and events in time, they are complex and stratified objects (Sassoon 1998) to which a plethora of different meanings can be attached, depending on the observer. Digitisation adds to the complexity of these objects by requiring them to be re-evaluated in the context of a project, even though the capture in digital format partially strips them of their material properties and the change in context affects their relational values (Note, 2019). Artificial Intelligence (AI) can help us enrich the intellectual properties we associate with digitised photographs by adding me- tadata, opening up new possibilities for item-level control and de- scription. Computer Vision (CV), part of the AI methodologies that include "activities that identify human faces and objects in digital images" (Samoili et al. 2020:13), has influenced the way visuality is understood (Smits / Wevers 2022) and even contribu- tes to the development of the field of "visual digital humanities", as identified by Münster and Terras (2020). The emerging interest of memory institutions in adopting AI, mainly to facilitate the production and use of their digitised collec- tions (Markus et al. 2021), is likely to lead to the outsourcing of ex- pertise in dealing with AI systems. Professional training has only recently begun to support the development of advanced computa- tional skills - as is happening in digital curation (Cushing / Kal- pana 2019). Furthermore, to date, the various curatorial approa- ches rooted in the habitual (and born-analogue) working practices of memory institutions (Hughes 2011) have informed the use and production of digital standards and data models. Is there a degree of conceptual and practical compatibility bet- ween existing curatorial practices in memory institutions and available CV workflows? This poster seeks to address this ques- tion by considering openly licensed digitised historical photo- graphs from the Irish context. It presents initial findings from the first author's PhD project, which explores the implications of ac- cessing and reusing digitised collections for CV tasks. A subset of the Photographic National Folklore Collection re- sulting from the University College Dublin digitisation project 1 and some of the thematic photosets from the National Library of Ireland's Flickr profile 2 were retrieved and explored using data mi- ning. Voxel51's FitfyOne 3 is the open source toolkit used to sup- port the workflows and facilitate the curation and visualisation of the dataset - as well as evaluating the performance of the CV mo- dels and libraries in use. The work is still in progress, but preliminary results suggest that the different archival descriptions provided for the collections ana- lysed natively offer different ways of creating access points to the photographs. The occasional incompleteness of the metadata and the lack of information on the pre-digitisation evaluation process limit the interpretability of the photographic collections and the- refore the potential CV applications. Notes 1. https://www.duchas.ie 2. https://www.flickr.com/photos/47290943@N03/ 3. https://voxel51.com/ Bibliography Barthes, Roland (1981): Camera lucida: Reflections on photo- graphy . Macmillan. Cushing, Amber / Shankar, Kalpana (2019): “Digital cura- tion on a small island: a study of professional education and trai- ning needs in Ireland”, in: Archives and Records 40, 2: 146–163. DOI: 10.1080/23257962.2018.1425135. Hughes, Lorna (2011): Evaluating and measuring the value, use and impact of digital collections . Facet Publishing. Markus, Gregory / Neudecker, Clemens / Isaac, Antoine / Bergel, Giles / Bailer, Werner / Marrero, Mónica / Tzouva- ras, Vassilis / Oomen, Johan / van Kemenade, Philo / Bontje, Marloes / Cuper, Mirijam / Bartholmei, Stephan / Cejudo, José Eduardo / Larsson, Albin / Angelaki, Georgia (2021): AI in relation to GLAMs Task Force. Report and recommendations . Europeana Network Association. https://pro.europeana.eu/pro- ject/ai-in-relation-to-glams. Münster, Sander / Terras, Melissa (2020): “The visual side of digital humanities: a survey on topics, researchers, and epistemic cultures”, in: Digital Scholarship in the Humanities 35, 2: 366– 389. DOI: 10.1093/llc/fqz022. Note, Margot (2019): “The Picture Edge: Archival Appraisal of Photographs”, in: Margot Note Consulting LLC . https://ww- w.margotnote.com/blog/2019/08/26/picture-edge [05/10/2022]. Samoili, Sofia / López Cobo, Montserrat / Gómez, Emi- lia / De Prato, Giuditta / Martínez-Plumed, Fernando / Deli- petrev, Blagoj (2020): AI watch: defining Artificial Intelligence towards an operational definition and taxonomy of artificial in- telligence . LU: Publications Office of the European Union. DOI: 10.2760/382730 Sassoon, Joanna (1998): "Photographic Meaning in the Age of Digital Reproduction", in: LASIE 29, 4: 5–15. Smits, Thomas / Wevers, Melvin (2022): „The agency of com- puter vision models as optical instruments“, in: Visual Communi- cation 21: 329–349. DOI: h10.1177/1470357221992097. 1