Social media has become a handy tool for people to share events when formal news portals have limited access, but sometimes it can be difficult to distinguish valid posts from rumored ones. In response to this growing phenomena, researchers are planning for a social media lie detector that will analyze the nature of online posts in real-time.
Drawing data from media blasts during the London riots in 2011, the project includes posts on Twitter, comments in healthcare forums and public comments on Facebook. By categorizing online rumors into different groups, researchers can more easily identify which rumors hold authority in relation to their source, which can range from journalists to internet bots. There will be four categories: Speculation, Controversy, Misinformation and Disinformation.
The system will analyze posts from many perspectives, as it will check an account’s history to see if it was created for a specific task and if other sources will provide contextual support online.
With the first set of results expected to launch in 18 months, researchers say they will create a “visual dashboard” for users to see if a rumor is taking hold, an effective tool for both journalists and governments to respond more effectively to events.