Putland, Emma and Chikodzore-Paterson, Chris and Brookes, Gavin (2026) ‘No fragrance of life, no essence of humanity’ : Critical Discourse Studies, Generative Artificial Intelligence, and Dementia. In: Critical Methodologies in Dementia Studies :. Routledge, London. (In Press)
Putland_et_al_CDS_GenAI_and_Dementia_Final_clean_Draft_May_26_.pdf - Accepted Version
Restricted to Repository staff only until 1 January 2050.
Available under License Creative Commons Attribution.
Download (1MB)
Abstract
Generative artificial intelligence (GenAI) is increasingly influencing how humans communicate about and understand all manner of topics in the world, including health-related ones like dementia. In this chapter, we consider GenAI’s potential to reproduce and amplify pre-existing stereotypes and harmful representations that have been found across human-produced writing and images (which GenAI has been trained on). We take a Critical Discourse Studies (CDS) approach, which is essentially concerned with examining how communicative choices both reflect and help to shape our worlds, here focusing on dementia. We present two case studies: (1) 171 images generated from the text prompt ‘dementia’ in Stable Diffusion (version 1.4); and (2) fifty-two AI-generated written descriptions of characters with dementia produced using Sudowrite, totaling 22,638 words. The chapter summarizes some of the key linguistic and visual patterns identified across the two datasets, which present restrictive stereotypes for people with dementia and foreground ableist/ageist discourses of dementia that may reflect or even amplify human representations. Tropes identified include a focus on pathology, loss (of self), suffering, social distance, and death. The chapter reflects on the implications of such representations, alongside some of the methodological challenges and opportunities associated with this topic.