A new study from the University of Aberdeen has revealed that artificial intelligence may be shaping the future of rewilding in ways that create an idealised version of what restored nature can look like.
For the study, which is published in People & Nature by the British Ecological Society, researchers analysed AI‑generated images and descriptions of rewilding produced by chatbots and compared them with visual and textual materials used by UK rewilding advocacy organisations. The findings show that both sources consistently favour familiar, idealised aesthetics: dramatic landscapes, charismatic wildlife, and scenes notably free of people or the “messy” ecological processes that underpin real ecosystems — including decay, disturbance, and death.
The study warns that by repeatedly showcasing such selective imagery, rewilding organisations are inadvertently training AI systems to reproduce exclusionary visions of nature. As AI tools become increasingly embedded in communications and public engagement, these biases risk being reinforced and widely disseminated, shaping how future audiences perceive rewilded landscapes.
However, the researchers emphasise that this trend can still be reversed. By incorporating a wider, more inclusive visual content — featuring diverse species, human presence, and the full spectrum of ecological processes — organisations can help steer AI models toward more socially accurate and ecologically authentic representations of rewilding.
Dr Flurina Wartmann from the University of Aberdeen said: “Our research highlights a pressing societal issue: conservation organisations are, often unintentionally, encoding incomplete visions of nature into the AI systems that will influence how the next generation imagines our environment in the future. It also demonstrates how social science methods can – and must be - used to critically interrogate AI‑generated content.
“This study marks one of the first examples of Human Geographers analysing GenAI‑produced visual materials and highlights the importance of ensuring that emerging technologies reflect inclusive, realistic, and equitable depictions of the natural world.”
This project was in part funded by the Leverhulme Centre for Nature Recovery, which is generously supported by the Leverhulme Trust.