Assistant Professor Kurt Luther receives National Science Foundation award
Assistant professor of computer science Kurt Luther has been recognized by the National Science Foundation with a Faculty Early Career Development Award to study and improve the capabilities of crowdsourced investigations.
The issue is of particular importance in an era where speed can sometimes best factual and accurate reporting of news.
Luther will use an innovative expert-led crowdsourcing approach to collect data using a platform called CrowdSleuth. The software will assist collaboration between crowds and experts, such as journalists, historians, and law enforcement, as they attempt to discover new information and verify details of investigations.
The project will specifically focus on investigations of photo and video evidence. Participants in the project will evaluate the effectiveness and efficiency of crowds in identifying names of people who appear in photos and in determining the time and place photos were actually taken. Participants will also analyze ethical issues surrounding the dissemination of information gleaned from the web.
“With so much information online, amateur and professional investigators have the ability to research very serious crimes, such as missing persons cases,” said Luther. “But even professionals have constraints on resources, like time and limits to their expertise. Some amateur investigators have proven quite useful in catching criminals. The flip side of crowdsourcing is that there is a growing concern over reckless digital vigilantism and investigations conducted online that can have disastrous outcomes for people in real life. The tragedy at the Boston Marathon in 2013 where online crowds identified and named innocent people as the perpetrators of the bombing is one unfortunate example of crowdsourcing gone wrong.”
Luther will also develop versions of his CrowdSleuth software for other domains, such as identifying historical photos in museum collections and investigating images of terrorist activity shared on social media. For each domain, he will interview experts and observe their work practices to understand their real-world needs and where crowdsourcing might help.
Luther hopes the research will transform the science of investigation. The goal would be to integrate methods from human-computer interaction and crowdsourcing that have mostly been overlooked, such as the collaboration of experts and citizen investigators, the development of a new software platform to facilitate that collaboration, and cultivation of a new sociotechnical way of thinking about how best to crowdsource information through these multi-pronged efforts.
“There is a need at this particular moment in time to research the nexus of scientific, technical, and social computing perspectives to ensure that the future of crowdsourced investigations can continue to grow as a force for societal good,” said Luther, who is also a fellow of the Institute for Creativity, Arts, and Technology. “This project is giving real tools to the public to make discoveries and solve mysteries with the help of an expert.”
The interdisciplinary nature of Luther’s project will be reflected in outreach activities that include co-teaching undergraduate courses and graduate seminars in investigative technologies open to students from areas as varied as computer science, multimedia journalism, history, and the Corps of Cadets who will become adept at building and using new investigative technologies. Beyond academia, Luther plans on hosting a Reddit Ask Me Anything event about his research and sleuthing workshops in the Washington, D.C., area.
Established in 1995, the NSF CAREER Award is the most prestigious award given by the NSF in support of junior faculty who demonstrate the potential to effectively integrate research and education. Luther’s grant is consecutive for five years and totals $554,628.
The award is the second grant Luther has received to study crowdsourcing in six months. He, along with professor of computer science T.M. Murali, is also currently studying the effectiveness of crowdsourcing in designing visual tools to study the effects of common chemical pollutants on cells.
Written by Amy Loeffler
