On the Historical Gaze of Generative AI: Visions of Scandinavia in Stable Diffusion
(2025) In Scandinavian Journal of History p.1-1- Abstract
- This article critically explores how the AI image model Stable Diffusion generates visual histories of Scandinavia in the 1930s. While the past has always been mediated, this study suggests that generative AI models constitute a black-boxed yet easily accessible tool for the narration of history. We draw on recent works in digital memory studies and use an explorative hands-on methodology to subject AI images to visual semiotic analysis to better grasp and discuss how generative models produce synthetic representations of the past. Focusing on how the model generates modernity symbols and historical figures, we unpack a series of issues concerning the relationship between AI and historical representation that warrants further research.... (More)
- This article critically explores how the AI image model Stable Diffusion generates visual histories of Scandinavia in the 1930s. While the past has always been mediated, this study suggests that generative AI models constitute a black-boxed yet easily accessible tool for the narration of history. We draw on recent works in digital memory studies and use an explorative hands-on methodology to subject AI images to visual semiotic analysis to better grasp and discuss how generative models produce synthetic representations of the past. Focusing on how the model generates modernity symbols and historical figures, we unpack a series of issues concerning the relationship between AI and historical representation that warrants further research. Along with the generated results, we discuss limitations in the training data, tension between generic past and historic specificities, as well as how Stable Diffusion produces an algorithmic tourist gaze, national stereotypes, and anachronisms. This article heeds the call to employ infrastructure-centric approaches to understand how tech companies govern users’ capacity to shape memory practices and craft popular images of the past. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/c5964d60-1bcc-4e96-a9fc-884cbfd3142b
- author
- Stjernholm, Emil LU ; Eriksson, Maria LU and Mohammadi Norén, Fredrik
- organization
- publishing date
- 2025-06-25
- type
- Contribution to journal
- publication status
- epub
- subject
- in
- Scandinavian Journal of History
- pages
- 31 pages
- publisher
- Routledge
- ISSN
- 1502-7716
- DOI
- 10.1080/03468755.2025.2511644
- project
- Modern Times 1936
- language
- English
- LU publication?
- yes
- id
- c5964d60-1bcc-4e96-a9fc-884cbfd3142b
- date added to LUP
- 2025-06-26 12:31:26
- date last changed
- 2025-06-26 13:01:03
@article{c5964d60-1bcc-4e96-a9fc-884cbfd3142b, abstract = {{This article critically explores how the AI image model Stable Diffusion generates visual histories of Scandinavia in the 1930s. While the past has always been mediated, this study suggests that generative AI models constitute a black-boxed yet easily accessible tool for the narration of history. We draw on recent works in digital memory studies and use an explorative hands-on methodology to subject AI images to visual semiotic analysis to better grasp and discuss how generative models produce synthetic representations of the past. Focusing on how the model generates modernity symbols and historical figures, we unpack a series of issues concerning the relationship between AI and historical representation that warrants further research. Along with the generated results, we discuss limitations in the training data, tension between generic past and historic specificities, as well as how Stable Diffusion produces an algorithmic tourist gaze, national stereotypes, and anachronisms. This article heeds the call to employ infrastructure-centric approaches to understand how tech companies govern users’ capacity to shape memory practices and craft popular images of the past.}}, author = {{Stjernholm, Emil and Eriksson, Maria and Mohammadi Norén, Fredrik}}, issn = {{1502-7716}}, language = {{eng}}, month = {{06}}, pages = {{1--1}}, publisher = {{Routledge}}, series = {{Scandinavian Journal of History}}, title = {{On the Historical Gaze of Generative AI: Visions of Scandinavia in Stable Diffusion}}, url = {{http://dx.doi.org/10.1080/03468755.2025.2511644}}, doi = {{10.1080/03468755.2025.2511644}}, year = {{2025}}, }