Digital nature in the AI era: How human and AI ‐generated representations shape future visions of rewilding

Abstract Rewilding has gained significant influence in nature conservation, offering hopeful narratives that address the interconnected challenges of climate change and biodiversity loss while enabling people to reconnect with ‘wildness’ in the Anthropocene. Rewilding powerfully shapes imaginaries of possible futures and is being implemented through numerous projects across the UK. These projects fundamentally transform landscape aesthetics and determine which human and non‐human species are rendered visible, desirable and legitimate in rewilded futures. In this study, we interrogate the aesthetics of rewilding through a qualitative approach that examines visual and textual materials from advocacy organisation websites, their social media accounts on Instagram and artificial intelligence (AI)‐generated representations of rewilding produced by AI chatbots. Our analysis focuses on the framing and content of visual materials, such as what is represented and what is absent, as well as how these materials construct visions of desirable future natures. We find that representations of rewilding focus on conventional nature aesthetics, predominantly excluding aesthetically challenging ecosystem processes, non‐charismatic species and a diversity of people. We argue that visual and textual materials reveal not only how nature is represented in rewilding discourses but also embody normative assumptions about what recovered or rewilded nature should look like. These visual representations act politically by normalising the presence or absence of particular people and activities in landscapes, thereby perpetuating narrow conceptualisations of environmental futures. We discuss how our findings may inform the development of more socially just environmental imaginaries, contributing to more inclusive future land management both in policy and practice. Read the free Plain Language Summary for this article on the Journal blog.