The Media Bias Taxonomy
In this project, we did an extensive literature review of papers related to media bias and automatic media bias detection [1]. We developed a categorization of different types of bias, as shown above, which is always a work in progress. The second iteration of the framework, complete with definitions, is on its way! We also reviewed all recent papers on media bias detection in computer science and concluded that while pre-trained and fine-tuned LLMs are the most commonly used methods, there is a wide range of promising alternatives.
There’s a blog post on the Media Bias Research Group website, if you prefer this format over the long paper (here’s the post on detection methods). At some point, I will explain the different kinds of bias with their definitions in easier language here. 👀