Artificial intelligence may transform the way people access information as they make decisions about nature-based solutions. This image was created by the authors with the assistance of DALL·E 2.

By Daniel Richards, David Worden, Xiao Ping Song, and Sandra Lavorel.

Read the full paper here.

Nature-based solutions are those that use ecosystems to sequester carbon, mitigate climate risks, and develop new economic opportunities and resilience (e.g. by planting trees or creating wetlands). The success of these diverse solutions is context dependent, requiring a breadth and depth of scientific expertise when providing information. The high cost and low availability of adaptation expertise constrain the uptake of nature-based solutions globally.

To adapt to climate change, stakeholders need locally tailored guidance about climate risks and the potential of nature-based solutions. It is currently expensive and time-consuming to provide this guidance. Generative artificial intelligence (GenAI) may transform the communication of adaptation guidance by reducing the cost and increasing the speed at which information can be provided. Specifically, we see potential for GenAI to write tailored reports for any land parcel automatically, provide instant responses to stakeholder queries at any time of the day or night, and create clear, accurate, and visually attractive images to communicate diverse future scenarios.  

This article highlights three worked examples of how GenAI technology could support decision-making about nature-based solutions. We used currently available and generic GenAI algorithms to demonstrate that there is presently significant potential in this field, which could be better realised by future technological development and research. First, we used GenAI to synthesise scientific information and write summary reports to guide future decisions about nature-based solutions on two farms. Second, we used GenAI to interactively provide real-time advice and respond to questions about the design of a private garden to support biodiversity. Third, we used GenAI to create visual representations of contrasting future land use scenarios for a given landscape. These examples highlight potential applications of GenAI, but we must also be mindful of the risks of this new technology. Notably, GenAI brings risks of data bias, false information, data privacy, mistrust, weak accountability, and inequity. Furthermore, the training and usage of GenAI models may have a substantial carbon footprint.