AI With an Impact: UMD Team Testing Bilingual Chatbot for New Moms

Feb 29, 2024

An AI-driven chatbot could soon be helping bleary-eyed moms of newborn babies manage the uncertainty of caring for their infants.

Jordan Boyd-Graber, an associate professor of computer science with an appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS), is part of an interdisciplinary team using AI to provide 24-hour, validated information to new moms.

The chatbot they developed, called Rosie, offers health information on topics ranging from medication dosages to post-partum depression to simpler queries like determining the proper temperature for bathwater or asking whether it’s normal for babies to lose their hair.

All the information provided is from reputable sources like the Centers for Disease Control and Prevention, Children’s National Hospital, Mayo Clinic, and more.

Boyd-Graber, working with Ph.D. student Neha Srikanth, is tasked with the computational side of the project, including optimizing the AI-powered question and answer software, and determining best practices for hosting the platform.

They are part of a UMD team that includes epidemiologists, biostatisticians, and behavioral and community health experts.

“One of the reasons I love this collaboration is that we have a process for collecting people who might be able to use the software, getting it in their hands, making sure they can use it, and then testing whether we actually made an impact,” Boyd-Graber says.

The project was initially funded by a $3.7 million award from the National Institutes of Health in 2021; it received an additional $200,000 last fall to expand Rosie’s outreach to Spanish-speaking mothers. The app is currently being tested in clinical trials.

In addition to his UMIACS appointment, Boyd-Graber is a core faculty member in the Institute for Trustworthy AI in Law & Society (TRAILS). Both he and Srikanth are active in the Computational Linguistics and Information Processing (CLIP) Laboratory, which is part of UMIACS.