This project addresses how factual claims and fact-checking efforts circulate across languages, and how generative AI may be reshaping this ecosystem through amplified claim production and AI hallucinations. Using the MultiClaim dataset, which contains approximately 200,000 claims and fact-checking metadata across 39 languages, I will analyze spatio-temporal patterns, cross-lingual claim diffusion, and gaps between misinformation spread and verification responses. Methodologically, the project will combine machine learning to identify AI-generated or AI-influenced claims, including fabricated citations, invented statistics, and non-existent entities. The project aims to develop the foundation for a larger research grant proposal, and generate practical insights for fact-checkers, journalists, and policymakers. A longer-term output is to inform AI-assisted multilingual verification systems that improve detection accuracy, robustness, and response speed.
Dieses Ereignis ist vorbei.