The role of artificial intelligence tools in climate change scientific assessments

Abstract

Scientific assessments, such as the IPCC reports, are vital tools for evidence-based decision-making. However, the exponential growth of scientific literature and the length of these reports can hinder the efficiency of the assessment processes, and the accessibility of their findings. This perspective article explores how existing and emerging artificial intelligence (AI) tools could offer solutions to these challenges. The article discusses the use of advanced information retrieval, text mining, and machine learning tools to navigate the increasingly vast climate change literature and support consensus-building. It argues that expert judgment remains crucial for assessing compiled literature. The article also looks at the potential of Large Language Models (LLMs) to communicate the complex findings of IPCC reports, emphasising the need to preserve scientific nuance, complexity and uncertainty. The article concludes with proposed hybrid ways forward for combining AI solutions with expert judgment could significantly improve the effectiveness of future IPCC assessments.

Keywords: Scientific assessments; consensus building; science communication; evidence synthesis; machine learning; large language models; experts’ judgement