UCL Dissertation: Epistemological framework for AI-generated images
[Research] [AI]
This dissertation examines the epistemic functions of images in the age of artificial intelligence (AI), focusing on the distinction between analogue photographs, digital images, and AI-generated images. It explores the challenges posed by AI-generated images to our understanding of reality and authenticity, emphasising the need for an interdisciplinary approach that combines philosophy, art criticism, media studies, and computer science. The dissertation explores how these images challenge the traditional epistemic functions of photographs and digital images. Employing a mixed-method approach, including extensive literature review and critical discourse analysis, this research scrutinises the technologies and algorithms underpinning AI image generation, focusing on their transparency and interpretability.
The main findings are firstly, that AI-generated images significantly challenge the traditional epistemic functions of photographs and digital images due to their lack of direct connection to reality and the opaque nature of their image-generation processes. Secondly, despite technological advancements, there is a critical need for educational frameworks that foster the public's critical eye and interpretation skills of AI-generated images
