University of North Carolina at Chapel Hill

AI Journalist

University of North Carolina at Chapel Hill is one of the 2018 winners of the Challenge Fund. See all the winners.


  • Steven King, Chief Innovation Officer and Assistant Professor, UNC School of Media and Journalism, Special Appointment with Kenan-Flagler Business School
  • Stacy Joines, IBM Fellow and CTO, Watson Implementations
  • Dan Spears, Editor, Wilmington Star-News
  • Stacie Hidek, Editor, MyReporter

Describe your project

The Star News has been collecting readers’ questions for five years, trying to answer as many as humanly possible. Can an AI-powered journalist provide helpful and essential news and information to a local community enabling a new revenue option for publishers?

We have been partnering with The Star-News in Wilmington, North Carolina, for two years helping them develop new, sustainable media products and to build this prototype by working directly with the MyReporter editors (a product of The Star-News). We partner closely with MyReporter, which takes questions from readers and assigns a reporter to answer it. This project would expand that offering and answer many more questions than they can now.

What is your experiment?

Students will work with local journalists to train an IBM Watson-powered AI robot to respond to the community’s questions in ways that engage users and provide valuable insights about the interests of the community, enabling journalists to better report on what readers want. Students will work embedded within The Star-News newsroom and lead the product development, taking it from a proof-of-concept to a launchable product. The following semester, students will test the idea and do user studies and make recommendations on how to improve the experience. Students will work directly with the developers to prioritize and create the best user experience for the target audience.

2019 update: During field testing of the prototypes, we know that the app must provide the correct answer to more than 85% of the questions or people will not use it. This has proven difficult to date because there is such a range of questions being asked in numerous disciplines. Our students have taken the user feedback from field testing and designed an updated version that addresses those.

If the experiment works, what do you think might happen?

Journalism publications across the country will integrate and adopt the AI Journalist to better inform their local communities, while students will have more opportunities to test and work in various markets.

How is this project unique and innovative?

The intelligence platform is built in IBM Watson with integrations into Facebook Messenger, as well as the local publisher’s website and mobile app. We have a unique relationship with the Watson team and have their full technical support on this project. The AI Journalist indexes numerous local sources of news and information and learns from previous search and responses to provide the most relevant information while also providing important information to the journalists covering the community. No other project can provide such relevant responses to seemingly infinite user questions helping to better engage and inform that community.

How might this experiment change teaching at your school or media practices in your partner’s newsroom?

Real-world, project-based learning is essential to engage and equip students for the modern and ever-changing media world. This opportunity will be part of the Storytelling in Emerging Technologies Class and the ET Lab, allowing students to work directly with both journalists and new technologies to solve difficult problems.

2019 update: The students are learning that the school and project has more resources than a local newsroom. The newsroom is providing great support to the students and the students are learning what it takes to produce a long-term project while the paper is getting produced daily.