AI and the Historian: Why Digital Literacy Matters Now More Than Ever

Journal of Victorian Culture's Digital Forum has been leading discussions on digital transformations in Victorian Studies for well over a decade.From 2016-2019, I co-edited a series of issues with Dr Christopher Donaldson.In a 2017 forum, 'Workflow' , Christopher and I cautioned readers about historians' increasing dependence on digital sources, in particular, an over-reliance on that great open sesame of digital archives: the keyword search. 1 Conscious that relying on digital collections could lead us to being less critically reflexive or forthright about our work, we encouraged scholars to open the 'black box' of digital archives and develop new forms of source criticism that acknowledged how original documents are manipulated, recoded, and reframed through the digitization and web-publication process.As Tim Hitchcock argued, ' Algorithm-driven discovery and misleading forms of search, poor OCR, and all the selection biases of a new edition of the Western print archive have changed how we research the past, and the underlying character of the object of study (inherited text)' . 2 Since that time, the introduction of new tools and techniques, notably machine learning and its subsets, including generative artificial intelligence models such as ChatGPT, means we need to renew existing forms of historical data criticism.As novel digital research infrastructures and tools emerge, how will this impact on our understanding of the nineteenth century and even on how we 'do' history?In the era of misinformation -fake news, conspiracy theories, and deepfakes -data literacy is an increasingly urgent issue for scholars, students and citizens alike.Even in relevant subject areas, like English and History, where digital humanities techniques are applicable and powerful, some academics still fail to recognize the importance of digital literacy to staff and students and the contribution that digital humanities can make.Yet the core competencies of the historians' toolkit -source criticism, provenance, historic contextualization, uncovering bias -are crucial for developing digital literacy, and Digital Victorianists are uniquely placed to contribute.This article reflects upon on my experiences co-developing a series of databases and web resources over the past decade.Most recently, I have been involved in two digital history projects: Data Mining Convict Tattoos (British Academy/ JISC, 2019) which examined the largest number of tattoos ever recorded: 75,688 descriptions of tattoos, on 58,002 convicts in Britain and Australia from 1793 to 1925, and Skin and Bone: Interdisciplinary Analysis of Accidents, Injury and Interpersonal Violence in London, 1760-1901 (British Academy/ Leverhulme, 2021-22) which merged convict, hospital and osteoarchaeological datasets and documented 87,903 injuries on 50,659 Londoners, revealing the physical impact of the Industrial Revolution on the body.These projects emerged from the Digital Panopticon (Arts and Humanities Research Council, 2014) which created a single, searchable database of 250,000 individuals sentenced at the Old Bailey between 1780 and 1868, and assessed the impact of varying modes of punishment on their lives.