You’ve got Bias - The news bias detector
In 2005 hurricane Katrina swept in over New Orleans causing massive destruction due to flooding and rendered people homeless. Media reporting on the event had a tendency to disproportionately associate black people with crime and violence. An image that has circulated shows a black man with a caption indicating that he is “looting” while a seemingly comparable white couple are described as “finding” food.
|Black man “looting” and white couple “finding” food|
News media also reported on events that never took place such as rapes and schools filled with hundreds of dead bodies. This is an extreme example of hidden bias influencing media but with the help of future technology we will make this bias visible to the public.
We will construct an automated system that analyzes a news article and uses a number of available data regarding the author as well as other articles of the same subject to point out hidden bias. For example, when analyzing a technology article about how good the latest Apple product is, our system may point out that all of the editorial staff uses Apple computers in their work. Sources we are considering includes conventional media articles, tweets and blog posts as well as data from public records such as age, nationality, residential history and company records. The system will show information about processes and influences that were key in the creation of the news-story, giving the audience not only the story but also some clues about how the story came into existence.
We see a future where everyone can be their own personal news channel. Given this scenario it will be more difficult to know who the sender is and misinformation due to bias will be spread more easily. Our hope is that this system will help make journalists more aware of their own personal bias and help readers navigate this new media landscape.
/Niklas Berg, Jakob Florell, Terese Nothnagel, Gustav Rannestig