Generative Artificial Intelligence and Digital Ageism: Exploring the Construction of Age and Aging by Image-Generating AI

Since 2022, the increased attention to and public accessibility of generative artificial intelligence (GenAI) has become essential for knowledge acquisition on digital platforms. GenAI for image generation is trained on large datasets but remains socially situated in the biases of its creators as well as the inherent limitations and biases of the training data used. Existing research has pointed out that AI models display gender and racial biases. However, the visual construction of age through GenAI models remains unexplored. This paper conducts a documentary image interpretation of 200 images of aged individuals to analyze the visual construction of age across different AI models. The results show different constructions of age between the models DALL·E and Stable Diffusion, ranging from an uncanny idealism to the fragility of age.

Medien & Altern [Media & Aging]