Climate change and its impacts extend far into the future; therefore, long-term perspectives are important for taking urgent climate action today and planning for both the short and long term. Long-term scenarios play a pivotal role in climate research, and are used to predict how climate change will be valued and how societies will be affected.

“AI is having a growing impact. Therefore, it is necessary to take this into account when studying climate policy issues. AI has the potential to change the conditions for how we conduct climate policy,” said Fredrik Heintz, professor at Linköping University and performing research in Mistra Geopolitics.

The development of AI has likely reached an inflection point, with significant implications for how research addresses emerging technologies and how it will drive long-term socioeconomic development linked to climate change scenarios. When researchers try to predict the future, they consider factors like economy, geopolitics, climate change, emerging technologies and many other factors, and AI will affect all of these areas and the role they play in long-term scenarios.

“As researchers scramble to grasp the development of AI, and the challenge of building long-term socioeconomic scenarios increases, AI could hold potential to aid this process. It cannot be ruled out that AI systems most likely will be better at predicting the future of AI than humans are at predicting the future of humans,” said Henrik Carlsen, Co-Director for Mistra Geopolitics and Senior Research Fellow at SEI.

Shared Socioeconomic Pathways (SSPs) are climate change scenarios of projected socioeconomic global changes up to 2100, as defined in the IPCC Sixth Assessment Report on climate change in 2021. They are used to derive greenhouse gas emissions scenarios with different climate policies.

The authors of the new commentary, Henrik Carlsen, Somya Joshi, Fredrik Heintz and Björn Nykvist, argue that AI already shapes societal development and might have outsized impacts during the timeframe of SSPs. Given that AI could impact all drivers in the SSPs, it has considerable potential to fundamentally change societies in ways important for research and policy.

Carlsen adds that he sees three possible tracks for the future use of AI in climate research:

1. We will stop predicting the future. It is very difficult and challenging to build long-term societal scenarios. The uncertainties might be so great for the long-term scenarios that we stop predicting the future. For example, climate research communities have been hesitant to pick up new scenarios (technological forecasting).
2. We will start using more systematic methods and flexible scenarios, for example a database with influencing factors, and build combinations of these factors. This approach would be more detailed and adapted to the context.
3. We will embrace AI to assist us when creating long-term scenarios. But it cannot be ruled out that humans are still humans.