Climate AI Nordics

Climate AI Nordics

Nordic network for collaborations around tackling climate change using AI and machine learning.

Inaugural Nordic Workshop on AI for Climate Change sparks collaboration and looks to the future

The first annual Nordic Workshop on AI for Climate Change concluded in Gothenburg on May 13, marking a significant and successful milestone for research on AI for climate action in the Nordic countries. The event was a vibrant convergence of leading researchers, practitioners, and innovators, all united by their work at the critical intersection of artificial intelligence and climate action. Throughout the day, the venue was alive with conversations and focused engagement. Computer vision specialists could be seen in discussion with marine biologists, while atmospheric chemists and soundscape analysis researchers shared perspectives around poster displays. Ideas found common ground.

Attendees explored a wide array of impactful topics, from harnessing citizen science and AI for more accurate species range estimation to deploying soundscape analysis and bioacoustic event detection for enhanced biodiversity monitoring. Discussions also delved into the crucial environmental footprint of AI itself, particularly within environmental monitoring, and showcased AI’s growing role in understanding atmospheric chemistry. The power of Large Language Models to extract critical insights from climate-related texts and to evaluate the reliability of specialized language models was another key theme. Furthermore, the workshop highlighted how AI, combined with Earth observation data, is revolutionizing our ability to scale up complex tasks such as detailed tree cover mapping and insightful socio-economic analyses in our rapidly changing climate.

We are immensely grateful to all the incredible speakers whose presentations sparked such engaging dialogue, and to every participant who contributed to the workshop’s dynamic and collaborative atmosphere. Our sincere thanks go to our partner network, Climes - Swedish centre for impacts of climate extremes, for their invaluable co-arrangement of the event. The keynote talks and oral presentations were amazing, setting a high bar for future events.

Building on this fantastic momentum, we are thrilled to announce that the 2026 Nordic Workshop on AI for Climate Change will be hosted by our core team members at the University of Copenhagen and the Pioneer Centre for AI, on June 26, 2026. Keep an eye out for more information on this and other community events!

We are immensely proud to be building this vital community. The success of our inaugural workshop is a testament to the collective passion and expertise in the Nordic region. Let’s continue to connect, collaborate, and tirelessly work towards maximizing the important potential of AI to address the urgent climate crisis facing our planet.

Published:

Categories:

You May Also Enjoy

Climate AI Nordics Newsletter, May 2025

Published:

Welcome to the Climate AI Nordics Newsletter, May 2025! Catch up on the Nordic Workshop on AI for Climate Change, major tree loss in India’s agroforestry, and upcoming events on conservation, causal inference, and digital emergency systems.

Severe decline in large farmland trees in India over the past decade

Published:

This study investigates changes in agroforestry tree cover across India between 2010 and 2022 using high-resolution satellite imagery (RapidEye and PlanetScope) and extensive tree label datasets. The researchers applied deep learning techniques from object counting, using heatmap-based models that transformed discrete tree labels into continuous Gaussian representations for improved detection. Their findings reveal a significant decline in large farmland trees, with up to 11% loss nationally and as high as 50% in certain hotspot areas. From 2018 to 2022 alone, over 5 million trees were lost, reflecting a rapid transformation in India's agricultural landscapes. Interviews with farmers linked this decline to shifts in cultivation practices, particularly the intensification of cropping systems.

High-stakes decisions from low-quality data: AI decision-making for conservation

Event date:

Webinar with Lily Xu, Columbia University. Like many of society's grand challenges, biodiversity conservation requires effectively allocating and managing our limited resources in the face of imperfect information. My research develops data-driven AI decision-making methods to do so, overcoming the messy data ubiquitous in these settings. Here, I’ll present technical advances in machine learning, reinforcement learning, and causal inference, addressing research questions that emerged from on-the-ground challenges in wildlife conservation. I’ll also discuss bridging the gap from research and practice, with anti-poaching field tests in Cambodia, field visits in Belize and Uganda, and large-scale deployment with SMART conservation software.

Debiasing AI predictions for causal inference without fresh ground truth data

Event date:

Webinar with Markus Pettersson, Chalmers University of Technology. Machine learning models trained on Earth observation data, particularly satellite imagery, have recently shown impressive performance in predicting household-level wealth indices, potentially addressing chronic data scarcity in global development research. While these predictions exhibit strong predictive power, they inherently suffer from shrinkage toward the mean, resulting in attenuated estimates of causal treatment effects and thus limiting their utility in policy evaluations. Existing debiasing methods, such as Prediction-Powered Inference (PPI), require additional fresh ground-truth data at the downstream causal inference stage, severely restricting their applicability in data-poor environments. In this paper, we introduce and rigorously evaluate two novel correction methods—linear calibration correction and Tweedie's correction—that substantially reduce prediction bias without relying on newly collected labeled data. Our methods operate on out-of-sample predictions from pre-trained models, treating these models as black-box functions. Linear calibration corrects bias through a straightforward linear transformation derived from held-out calibration data, while Tweedie's correction leverages empirical Bayes principles to directly address shrinkage-induced biases by exploiting score functions derived from predicted outcomes. Through analytical exercises and experiments using Demographic and Health Survey (DHS) data, we demonstrate that both proposed methods outperform existing data-free approaches, can achieve significant reductions in attenuation bias and thus providing more accurate, actionable, and policy-relevant estimates. Our approach represents a generalizable, lightweight toolkit that enhances the reliability of causal inference when direct outcome measures are limited or unavailable.