Researcher Enhances Tool to Help Users Detect Misleading Headlines, Videos
By Annie Dankelson | Illustration by LAUREN BIAGINI
Five things you need to know! You won’t believe what happened next! Why we love ... misleading headlines!
While they aren’t always malicious, they can lure the absentminded social media scroller to click onto irrelevant content or, worse, an internet scam. Now, with a boost from a $228,000 National Science Foundation grant, a UMD researcher is developing a computational tool to help users identify and avoid such clickbait in both text and video.
“We want to automatically detect these things so that our (online) experience becomes better and our social network becomes more secure,” says Naeemul Hassan, an assistant professor of journalism and information studies and affiliate assistant professor of computer science.
Hassan is building on his earlier work on BaitBuster, a browser extension that uses machine learning to determine whether a text headline corresponds to its article’s actual content, then gives Facebook users a visual warning if it doesn’t. BaitBuster 2.0 will expand to combat video clickbait, requiring new algorithms to compress clips, gather their images and process their transcripts.
To increase the impact of the revamped BaitBuster, Hassan and his team will also organize training workshops and coordinate outreach efforts to help engage underrepresented groups in cybersecurity.
