Age does matter: Midjourney’s visual representation of older adults and journalistic ethics

Abstract

With the emergence of Artificial Intelligence (AI), the journalistic and informational media system is undergoing a profound transformation that significantly challenges the production and distribution of information. AI, particularly Generative AI (GenAI), poses risks and potential harms to society, reinforcing stereotypes based on gender, race, origin, and age. This article explores how ageism manifests in GenAI as an example of how AI reinforces stereotypes and its implications for media professionals. We focus on the visual analysis of images generated with Midjourney, one of the most popular AI image-generative tools. We defined 91 prompts describing everyday activities and combined them with two profiles (“person” and “older person”) to produce 1,456 images, along with their automatically generated descriptions. These images were then quantitatively analysed to examine demographic characteristics and representations of age. Our results show that most images representing an age-neutral person depict a young individual, while older adults are scarce and relegated to secondary roles and negative depictions. Older individuals are portrayed with signals of vulnerability, fragility, and dependence, while gender and ethnic stereotypes become more extreme when associated with age. GenAI ageist depictions represent a step back in the more diverse and inclusive vision that the legacy media has adopted over recent decades regarding age. GenAI offers media professionals opportunities to enhance productivity, but addressing its biases is essential in news production; being critical of the images generated and disseminated to ensure fairness, inclusivity, and diversity has become an ethical imperative in journalism today, more than ever.