The U.S. intelligence community relies on anticipatory thinking, visualizing what might happen based on relevant known facts, to provide critical insight during the intelligence analysis process. Providing a diverse collection of training examples — stories where imagined bad actors form and pursue potentially nefarious goals — is both knowledge-intensive and time-consuming. However, systems to produce customized scenarios that are automatically designed and tailored to a trainee’s learning objectives may be on the horizon.

A grant from NC State University’s Laboratory for Analytic Science is supporting games researchers at the University of Utah who are working to develop a program for generating customizable narrative scenarios that support the analyst training process around anticipatory thinking. The work leverages ongoing research around the automatic generation of quests and storylines for video games.

“This project will make advances in the modeling of stories by computers, in order to create personalized training scenarios by machine. The scenarios will help intelligence analysts practice anticipatory thinking, a skill they need to quickly infer potential outcomes that might unfold from current events.,” says Michael Young, deputy director of the U’s Entertainment Arts and Engineering program and a professor of computer science at Utah, and principal investigator on the project.

Initially, the researchers will focus on having systems generate scenarios to around narco-terrorism. The system will build stories specifically designed to prompt anticipatory thinking about the likely actions of a drug cartel as its members pursue their goals. To that end, the researchers are collaborating with cognitive psychologists at Northern Illinois University to determine how to model goal-directed inferences made during story comprehension. The team also partners with intelligence experts at NC State’s Laboratory for Analytic Sciences.

The grant, titled “Anticipatory Thinking in the Context of Character Plan and Goal Dynamics,” will run for one year, from January 1, 2018 to December 31, 2018, and provides $130,000 to support the work.