Digital Doubles: How does AI-generated imagery affect human engagement with biodiversity conservation?

Abstract

1.Wildlife images created by generative artificial intelligence (AI) tools are rapidly growing in quantity and sophistication. Viewing these AI images could influence how people perceive and understand the real natural world, affecting their reactions to real wildlife imagery and potentially in direct wildlife interactions. 2.In this study, we look at how people donate money for conservation purposes to real and AI-generated birds, as well as whether they can accurately discriminate between images in these categories. We further investigate how participant moderating factors and perceived artificiality (i.e. how much an image looks AI-generated) affect these outcomes. 3.Using the contingent valuation method to assess willingness-to-donate, we found that participants award AI-generated and real bird images donation amounts that are not statistically discernible. However, greater perceived artificiality of an image was associated with higher donation amounts. No demographic factors significantly influenced donation behaviour.4.We also found that participants were able to correctly identify AI vs real images. A nature relatedness score measuring how connected participants were to nature being the only factor to positively predict ability to accurately choose between real and AI images. 5.These findings indicate the potential risk of AI-generated imagery replacing real nature imagery and experience, with artificial representations of wildlife being valued equally to real biodiversity. Conservationists must focus on promoting nature relatedness and real-life, direct nature experiences to combat AI-generated misinformation and decreased understanding of ecological realities.