AI-Enhanced Exhibits that Improve STEM Learning and Engagement

Intelligent Science Exhibits use mixed-reality technology and safely controlled AI to support children's STEM learning through hands-on scientific inquiry. Prior work has shown that AI-assisted physical experimentation significantly improved children's understanding of STEM concepts and engineering practices while increasing engagement, with visitors spending four times longer at AI-supported exhibits than at equivalent exhibits without AI. Building on these findings, this three-year interdisciplinary project investigates how conversational AI and speech recognition can further enhance inquiry-based learning in informal STEM environments. The project will examine how AI designed around evidence-based pedagogical principles can foster productive scientific conversations among children and caregivers while supporting engagement, collaboration, critical thinking, and learning across different STEM topics. This project is led by the Human-Computer Interaction Institute at Carnegie Mellon University in partnership with multiple museums and after-school organizations: Children's Museum of Atlanta, Children's Discovery Museum of San Jose, Muse Knoxville, MOXI, The Wolf Museum of Exploration and Innovation in Santa Barbara, Griffin Museum of Science and Industry in Chicago, and Boys & Girls Clubs of Western Pennsylvania. These partnerships provide a range of informal learning settings that enable evaluation of the needs of different audiences and assess whether design principles generalize across institutions and populations. Building on existing AI systems that use computer vision to provide real-time feedback during hands-on science experiments, the project will integrate speech-based interaction and scaffolding informed by research in the learning sciences and intelligent tutoring systems. The project will investigate how these capabilities influence scientific inquiry, caregiver-child dialogue, and STEM learning across multiple content areas, including Projectiles & Simple Machines, while extending our existing platforms on Stability & Balance, Forces & Motion, and Levers & Proportional Reasoning. Using user-centered design, formative evaluation, and controlled experimentation, the research will identify design principles for AI-enhanced mixed-reality exhibits that effectively support learners, caregivers, educators, and museum staff. Data collected across diverse informal learning environments will advance understanding of how safely controlled AI can improve learning, engagement, collaboration, and productive dialogue for children and families, while identifying design principles that generalize across content areas and contexts. This project was funded by the STEM K-12 program, which supports fundamental, applied, and translational research on STEM learning in both formal K-12 education and informal learning environments. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.