LiT.RL team meets with A.J. Wineck, a satirical writer who started the Western Beet. [Left to right: Sarah Cornwell, Yimin Chen, Victoria Rubin, A.J. Wineck, and Niall Conroy]
Since the summer of 2015 our Deception Detection team has been working on identifying satirical (or fake) news in the stream of legitimate news. In April 2016 we met with A.J. Wineck, a fifth-year economics student, who in fall 2015 had started a local satirical news website, the Western Beet. His writing is inspired by the Onion but he draws primarily on the local campus events (or non-events). The Western Beet satire, whether by A.J. or other contributors, typically includes a headlines, a stock photo, and a brief article. One particular article (see the full article or image and caption below) has drawn quite a bit of attention as well as indignation in the comments on the Facebook page, as its content was misunderstood by some of the readers to be a legitimate news.
It appeared that the context of the news parody was somehow lost on those readers. We were curious about A.J. perspective and asked him to come in an talk to us as a group who focuses on identifying this genre of news algorithmic ally. A.J. seemed to be quite surprised by the attention and said that the imagery may have created such resonance.
Our conversation with A.J. was amicable and informative. A.J. shared with us a few details on his creative process and his original and talked about how he started inspiration from a course he took with Dr. John Vanderheide at Huron College.
A.J. has graciously agreed that our lab could use the stories from his website as our data source for modeling and identification. Our algorithm for satirical news detection, trained on the Onion and the Beaverton stories, was able to pick up satirical news with about 84% accuracy. In a quick test of 10 sample stories from the Western Beet, we were able to automatically identify 9 of them as fake news, which is not entirely surprising, given that the Western Beet emulates the style of the Onion.