A Gamified Approach To Automatically Detect Biased Wording And Train Critical Reading
Smilla Hinterreiter · 2021 International Conference on Data Mining Workshops (ICDMW) · 2021
Biased media has an effect on the public perception of occurring events. By altering word choice, outlets can alter beliefs and views. A gold standard data set is needed to train sufficient classifiers that detect biased wording. This work aims to develop a game that trains players to read news critically while collecting their annotations. The vision is to tackle the complex problem of media bias detection with a very scalable, high quality, and gold standard data set to overcome the drawbacks of current models in the area.