Unsustainability Decreases Acceptance of Environmental Artificial Intelligence

Abstract: Climate change and environmental issues significantly impact human health and well-being, posing substantial challenges to global sustainable development. Addressing these challenges necessitates both immediate and long-term solutions. While climate change itself does not inherently elicit a moral response, potentially hindering public engagement in climate action, artificial intelligence (AI)—encompassing robots, algorithms, and models—emerges as a promising ally. AI’s ability to learn from experience, adapt to new inputs, and perform human-like tasks positions it as a critical tool in addressing environmental challenges.
However, AI presents a dual-edged impact on the environment. While it can support ecological governance and promote sustainability, its energy-intensive nature and associated carbon emissions may undermine environmental health. This study investigates the environmental implications of AI, focusing on how perceptions of AI’s sustainability influence public acceptance and the underlying psychological mechanisms.
Drawing on prior research, we hypothesized that individuals tend to avoid using unsustainable AI due to perceptions of diminished morality and heightened dependency (i.e., lower perceived agency). To test this, we conducted five studies employing diverse methodologies and measures. A preliminary study used 14 words generated by ChatGPT to gauge public attitudes toward AI in environmental decision-making, revealing a generally favorable view of AI in environmental management. Studies 1a and 1b experimentally manipulated perceptions of AI’s unsustainability, demonstrating that awareness of its high energy consumption and carbon emissions significantly reduces acceptance. Study 2 replicated these findings and identified morality—rather than dependency—as the mediating factor. Study 3 further revealed that pro-environmental attitudes significantly moderate the relationship between AI sustainability and its acceptance in environmental contexts.
In summary, this research highlights that while individuals are willing to collaborate with AI to address environmental challenges, their acceptance diminishes when AI is perceived as environmentally harmful. These findings underscore the critical importance of AI’s sustainability in achieving global sustainable development goals.