Data concerns undermine AI applications for climate.

The global climate crisis demands innovative solutions, and artificial intelligence (AI), including large language models (LLMs), offers significant potential to improve efficiency and enable lower-carbon lifestyles. However, public concerns around data privacy may affect the use of AI-enabled, data-driven applications with potential climate benefits. Drawing on three empirical studies with a nationally representative UK sample (N = 2078), we show that heightened awareness of AI's data collection practices reduces willingness to engage with such technologies - especially when perceived user benefits are low. We also find that data protection behaviours and greater familiarity with AI are associated with lower perceived risk, suggesting adaptive responses rather than disengagement. These behavioural dynamics highlight the importance of transparency, user agency and the integration of human and environmental dimensions into existing AI governance frameworks. Our results illustrate how public perceptions and everyday practices influence the social acceptance of digital innovations for sustainability.