The term media bias summarizes differences in news coverage about the same event, e.g., regarding content, tone, or perspective. Media bias can result from interventions of interest groups or unconscious biases of writers or readers. Reading news from diverse sources, e.g., from different countries, can reduce the effects of media bias. Although news is increasingly available online and often free, many people still read news from one or few sources, and thus are subject to media bias.News aggregators such as Google News are designed to work with thousands of sources. They crawl articles from online newspapers, group, summarize, and visualize related articles. However, due to limitations of current text analysis methods, also called NLP methods, no existing system can detect and reduce media bias. We develop an approach called broadened news analysis that is based on the concept of existing news aggregators. In contrast, however, our approach additionally reveals content-wise differences between related articles by sub-grouping them, e.g., by countries involved in an international conflict.
