Gaze-Based Fake News Detection: Insights from the FakeNewsPerception Dataset

Fake news remains a bane on public trust in information and reality, pushing gaze-based detection to take a different, human-centered approach, eye-tracking data used for the analysis of user interaction with misinformation. Insights from FakeNewsPerception-a framework that looks at gaze behavior, fixation duration, and saccadic behavior-are explored to discriminate between truthful and false content. Findings show how visual attention and cognitive biases exert influence over fake news perceptions-an alternative to conventional content-based approaches. Introducing gaze-based techniques into machine learning models shows the system improves accuracy and efficiency in fake news detection. These results bring forth wider horizons for building solutions aimed at combating misinformation and enabling credible online discourse.