===
What Happened
Artificial intelligence (AI) is advancing rapidly and is increasingly being used by scientists to better understand climate change. AI is perhaps best‑known for its use in “large language models”, such as ChatGPT, Claude and Gemini. From weather forecasting to analysing satellite images, climate scientists are using various types of AI to develop faster and more efficient ways to study the warming planet. However, the technology comes with no shortage of risks and challenges, including overreliance on unverified results, intensive energy use and exacerbating existing inequalities in climate science.
Carbon Brief spoke to 10 experts working across a range of climate‑related fields – from developing climate forecasts to analysing changes in the energy grid – about how AI is being used. Their answers are set out below, grouped by theme.
Benefits and Challenges of AI in Weather and Climate Forecasting
Prof William Collins, professor of AI for climate science at the University of East Anglia, said: “We’re seeing a really exciting renaissance in both weather and climate science… Machine learning has had an enormous impact on numerical weather prediction. AI systems have become good enough that they’re absolutely competitive with the world’s leading numerical weather prediction.”
Collins highlighted the creation of a “weather prediction model intercomparison project” that brings together physics‑based numerical weather prediction models, hybrid models and AI weather prediction models. He noted that the Intergovernmental Panel on Climate Change (IPCC) is performing its seventh assessment report (AR7) and, for the very first time, has a model intercomparison project just for models using AI. He described this as a real milestone and a harbinger of many more developments to come.
He added that AI should be telling us the known unknowns, not the known knowns, and that the value added by AI in the assessment process is to highlight uncertainty and gaps in knowledge.
Other Applications of AI in Climate Science
AI is also being used to analyse satellite imagery, hunting invisible methane leaks and detecting illegal logging, land‑use change and glacier retreat. Machine‑learning models process thousands of satellite images daily to identify these changes in near‑real time.
In polar research, deep‑learning systems have shown the potential to outperform leading physical models in seasonal Arctic sea‑ice forecasts, raising questions about the balance between model performance and the need for continued investment in observations.
Risks and Challenges
Experts warned that the rapid adoption of AI brings risks such as overreliance on unverified results, intensive energy use, and the potential to exacerbate existing inequalities in climate science. They emphasized the importance of maintaining robust observation budgets and independent validation to ensure AI tools remain trustworthy.
What This Means for Stakeholders
For planners, insurers, utilities and researchers, the key takeaway is that AI can offer faster, cheaper forecasts, but it must be used in a hybrid approach that preserves physical understanding and is validated against established models. Observation funding and independent evaluation remain critical safeguards.
Key Takeaway
- AI weather‑prediction models are competitive with leading numerical forecasts at lower computational cost.
- Experts favour hybrid approaches that combine AI with physics‑based models.
- The IPCC’s AR7 includes the first AI‑specific model intercomparison project.
- Maintaining observation budgets and independent validation is essential.
Leave a Reply