Innovative AI systems for nurturing and assessing creativity in k-12 learning environments

The National Science Foundation awarded the University of Denver $898,326 under the STEM Education program (CFDA 47.076) on September 15, 2025, to develop and evaluate two artificial intelligence systems for assessing and nurturing student creativity in K-12 learning environments.

The project delivers OCSAI Prime, a task-agnostic creativity assessment model using supervised fine-tuning to evaluate originality and appropriateness across diverse creative challenges, and Creativity Companion, an interactive facilitation tool employing reinforcement learning with human feedback to guide students through creative ideation in real time. Both systems will be made freely available on the web and evaluated through randomized controlled trials in middle school classrooms across three geographically diverse districts.

Work is performed in Denver, Colorado, with a period of performance through August 31, 2028.

The University of North Texas serves as a sub-recipient, with responsibility for preparing creativity assessment items and tasks, collecting data from schools, training graduate assistants for human scoring to support automated artificial intelligence scoring, conducting full implementation data collection for the Creativity Companion, and disseminating project outcomes. The University of North Texas sub-award was obligated January 21, 2026. A second sub-recipient, identified through a sub-award obligated March 19, 2026, will lead evaluation of both systems, contribute to technical development, mentor graduate students, and support project management and dissemination.

Abstract

Creative thinking is fundamental to science, technology, engineering, and mathematics (stem) success, yet current methods for assessing and nurturing student creativity are costly, time-intensive, and often inaccessible to under-resourced schools. This project addresses this critical national need by developing artificial intelligence-powered tools to make creativity assessment and mentoring both scalable and broadly accessible in k-12 education. The research team will create and evaluate two interconnected systems that evaluate student creative output and provide real-time guidance during problem-solving activities. By making these tools freely available on the web, the project will enable educators nationwide to systematically monitor and enhance student creativity. As artificial intelligence increasingly becomes part of professional creative work, the products from this work will also prepare students for a transformed future where human-machine collaboration in creative tasks will be commonplace. The project will develop two core artificial intelligence systems. The first, ocsai prime, represents a task-agnostic creativity assessment model that uses supervised fine-tuning to evaluate both originality and appropriateness across diverse creative challenges, moving beyond current tools that are limited to specific tasks. The second, creativity companion, employs reinforcement learning with human feedback to create an interactive facilitation tool that guides students through creative ideation in real-time. The companion will be trained using reward modeling based on ocsai prime's assessments combined with human feedback to optimize its ability to enhance student creative thinking. Both systems will undergo rigorous evaluation through a randomized controlled trial design in middle school classrooms across three geographically diverse districts, comparing student creativity outcomes, engagement, and learning persistence between artificial intelligence-assisted and control groups. The research will include comprehensive demographic analysis to ensure broad benefits while contributing new knowledge about computational creativity assessment, human-artificial intelligence CO-creativity, and educational technology integration in stem learning environments. This project is funded by the research on innovative technologies for enhanced learning (ritel) program that supports early-stage exploratory research in emerging technologies for teaching and learning. 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.- subawards are planned for this award.