LATET ANGUIS IN HERBA: UNVEILING AGEISM OF GENERATIVE AI

Resumen

Generative Artificial Intelligence (GenAI) burst into media industries as a cheaper, easier, and faster alternative to produce texts, images and videos. Social scientists are already starting to show how GenAI reinforces stereotypes and discrimination, partly due to gender, race and origin biases in the data used for machine learning. However, less has been said about ageism in GenAI. This is probably because ageist depictions are the norm, and ageism is widely disregarded and deprioritised in research, industry, and society. Building on ageist depictions, the media industry might be reinforcing stereotypes and harming self-perception. In this exploratory research, we analyse whether GenAI produces significant ageist outputs (or not) and to what extent GenAI reinforces age stereotypes through the analysis of 808 images generated with Midjourney. We compared older people's depictions with those of the general population. We employed computer vision and visual analysis to assess if Gen AI falls into associating ageist prejudices with older people. Preliminary results show that generative depictions not only fall into age stereotyping but also exploit mainly harmful tropes, disregarding all the diverse experiences of ageing. Thus, while the media industry has been moving towards an inclusive discourse, the ease of GenAI has burst as a threat to the inclusion of diverse perspectives.