The first ClimateCheck shared task at SDP 2025 challenged participants to build systems for scientific fact-checking of climate claims, with tasks focusing on abstract retrieval and claim verification against a corpus of 394,269 climate science publications.
Building on the success of ClimateCheck 2025, the 2026 edition expands to include disinformation narrative classification using the CARDS taxonomy, with substantially more training data and a focus on sustainable, real-world NLP systems.
A new study demonstrates that fine-tuned BERT models achieve competitive results with large language models for climate claim verification while processing claims 11-380 times faster, offering a more sustainable and practical solution for real-time fact-checking.
Building a collaborative research community to advance AI-powered climate and political fact-checking through interdisciplinary partnerships and open science.
