"Successful aging" with generative AI: cultural and ethical challenges in representing aging through synthetic...

Synthetic imagery produced by Generative Artificial Intelligence (GenAI) is rapidly transforming how aging is visually represented and culturally understood. However, the accuracy and representational adequacy of these outputs require further examination, especially when addressing minority or discipline-specific concepts, such as successful aging. This study critically examines GenAI imagery in East Asian contexts, focusing on how representations of gender, class, and successful aging are constructed and circulated. Drawing on the updated successful aging concepts, we employ visual semiotics as an analytical framework, situated within cultural gerontology, to treat synthetic images as algorithmic artifacts that reflect and reinforce normative ideals of later life. Using critical visual experimentation, we used Midjourney to generate 288 images for analysis, guided by prompts addressing social, affective, and environmental dimensions of aging. A total of 36 textual prompts were designed and grouped into three thematic categories: minimal descriptive prompts, scene-based prompts, and value-coded prompts. Our findings reveal gendered aesthetic norms, class-coded relational tropes, and algorithmic failures tied to ageist language. Images generated with prompts containing the phrases "successful aging" and "aging successfully" suggest that the image-generating tool embedded with natural language processing does not adequately interpret these expressions. We argue that GenAI compresses the complexity of aging into stylized, stereotypical, and reductive forms that risk amplifying cultural biases. At the crossroads of achieving reframing aging, our study calls for age-inclusive, culturally grounded AI design and contributes to interdisciplinary debates on synthetic visuality, representational justice, and AI-mediated culture, helping to advance a more inclusive and nuanced account of aging.