An Automated Approach to Fact Checking User-Generated Contentfor Journalists

Abstract: Journalists have exhibited widespread caution in regard to using content generated by citizen journalists because of the threat it poses to traditional news values and their role as gatekeepers. But text, photos and videos provided by citizens who witness breaking news events is one category of user-generated content that journalists are willing to use, which helps supplement their storytelling when fewer resources means journalists cannot personally be there. This willingness does not dampen several of their concerns; journalists require fact checking to be an integral part of any system that aggregates and accepts submissions by citizen journalists. This paper explores three different sets of tools that could be used to automate fact checking in a system that hosts user-generated content to ensure that content is accurate, neutral, timely and useful. These tools include vandalism detection software that identifies bias, inappropriate language, spam and copyright violations; crowdsourcing, which ensures that content is relevant, helpful and accurate; and photo and video authentication tools.