EBU LYNX tool : Liguistic property analysis for fake news detection

Presenter(s): Alexandre Rouxel, Pierre Fouché (EBU)

The EBU is exploring AI models using NLP and deep learning to assign reliability scores to texts, based on the the assumption that fake news publishers often have malicious intent, aim to influence large communities, and that this correlates with a particular writing style. The tool can be used to complement fact checking and provide input to journalists on the content they produce.