Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Myth-making machines - Analysing generative AI’s visual stories of climate change

Darge, Lars Jonas LU (2025) In Master Thesis Series in Environmental Studies and Sustainability Science MESM02 20251
LUCSUS (Lund University Centre for Sustainability Studies)
Abstract
Despite generative Artificial Intelligence’s (AI) explosive growth in popularity and record of biased outputs, little attention has been paid to the technology’s potential to shape perceptions around sustainability. This analysis investigates which perceptions about climate change causes, impacts and solutions are visually naturalised or marginalised by three text-to-image AI models: GPT‑4o, Ideogram, and Imagen. Guided by Barthes’ conceptualisation of semiotics and myth, and integrating content and semiotic analysis, a dataset of 960 images was coded for recurring visual elements. The emerging patterns revealed the pictures to tell six distinct myths, largely aligning with an ecological modernisation approach. A second step of analysis... (More)
Despite generative Artificial Intelligence’s (AI) explosive growth in popularity and record of biased outputs, little attention has been paid to the technology’s potential to shape perceptions around sustainability. This analysis investigates which perceptions about climate change causes, impacts and solutions are visually naturalised or marginalised by three text-to-image AI models: GPT‑4o, Ideogram, and Imagen. Guided by Barthes’ conceptualisation of semiotics and myth, and integrating content and semiotic analysis, a dataset of 960 images was coded for recurring visual elements. The emerging patterns revealed the pictures to tell six distinct myths, largely aligning with an ecological modernisation approach. A second step of analysis focusing on counter-narratives revealed only GPT‑4o to be able to create outputs strongly challenging the mainstream. These findings suggest a risk of generative AI hindering sustainability transformations by contributing to a visual climate change discourse that naturalises insufficient mainstream approaches and suppresses urgently needed alternatives. Solutions are explored. (Less)
Please use this url to cite or link to this publication:
author
Darge, Lars Jonas LU
supervisor
organization
course
MESM02 20251
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Visual climate change discourse, generative Artificial Intelligence (AI), myth, counter-narrative, Sustainability Science
publication/series
Master Thesis Series in Environmental Studies and Sustainability Science
report number
2025:038
language
English
id
9204565
date added to LUP
2025-06-24 10:06:17
date last changed
2025-06-24 10:06:17
@misc{9204565,
  abstract     = {{Despite generative Artificial Intelligence’s (AI) explosive growth in popularity and record of biased outputs, little attention has been paid to the technology’s potential to shape perceptions around sustainability. This analysis investigates which perceptions about climate change causes, impacts and solutions are visually naturalised or marginalised by three text-to-image AI models: GPT‑4o, Ideogram, and Imagen. Guided by Barthes’ conceptualisation of semiotics and myth, and integrating content and semiotic analysis, a dataset of 960 images was coded for recurring visual elements. The emerging patterns revealed the pictures to tell six distinct myths, largely aligning with an ecological modernisation approach. A second step of analysis focusing on counter-narratives revealed only GPT‑4o to be able to create outputs strongly challenging the mainstream. These findings suggest a risk of generative AI hindering sustainability transformations by contributing to a visual climate change discourse that naturalises insufficient mainstream approaches and suppresses urgently needed alternatives. Solutions are explored.}},
  author       = {{Darge, Lars Jonas}},
  language     = {{eng}},
  note         = {{Student Paper}},
  series       = {{Master Thesis Series in Environmental Studies and Sustainability Science}},
  title        = {{Myth-making machines - Analysing generative AI’s visual stories of climate change}},
  year         = {{2025}},
}