News
Research: how AI scrutinised more than 30 years of IPCC reports
Christophe Bisson, Associate Professor, member of the SKEMA Centre for Artificial Intelligence (SCAI) and Scientific Director of the MSc Business Consulting and Decision Intelligence in Paris, has co-authored a study examining how climate scenarios have evolved across the six synthesis reports published by the IPCC since 1990. A specialist in anticipatory systems, strategic intelligence and foresight, he used artificial intelligence to compare more than three decades of climate knowledge.
How AI reveals the evolution of IPCC climate scenarios
A study co-authored by Christophe Bisson, Associate Professor at SKEMA Business School, uses artificial intelligence to analyse how climate scenarios have evolved across the six synthesis reports published by the IPCC since 1990. Published in the open-access scientific journal PLOS Climate (Rank 1), the study was conducted by Christophe Bisson and Thierry Warin, Professor at HEC Montréal.
The two researchers designed a framework based on retrieval-augmented generation, or RAG, combined with a series of analytical prompts. This method enabled them to apply the same questions to all six synthesis reports published by the Intergovernmental Panel on Climate Change.
The analysis covers four areas: the relationship between mitigation and adaptation; the development of emerging technologies; socio-economic risks and pressures on governance; and the role of carbon dioxide removal in climate pathways.
Climate scenarios have become more precise
The findings reveal a clear continuity in the IPCC’s conclusions. Every report shows that delays in reducing emissions increase future risks and that the potential of technologies depends on public policy, funding, infrastructure and institutions.
The main change lies in the way this knowledge is represented. The earliest reports relied on qualitative reasoning. Recent publications incorporate more quantitative data, uncertainty levels, socio-economic pathways, and constraints linked to adaptation and the achievement of net zero.
“The aim is not to ask artificial intelligence to provide climate expertise. We use it to organise sources, apply the same questions to each report and make comparisons more meaningful. The method remains a tool that supports scientific judgement,” explains Christophe Bisson.
