ECR Projects

Explore past and current fundamental STEM education research projects across the three research areas that NSF's EDU Core Research (ECR) program funds, as well as across ECR funding types. Other search filters draw from both NSF's data and the ECR Hub's hand coding of award abstracts.

Ninth-grade biology students create cell models using clay.

Home > ECR Projects Search > Project Detail
STEM Learning and Learning Environments STEM Learning and Learning Environments  

Towards Embodied Learning for K-12 Machine Learning (ML) Education

Effective Years: 2023-2028

Children are increasingly impacted by technological advances in artificial intelligence (AI) that provide personalized recommendations regarding the books they read, individualized lessons, career-path plans, and friend circles. As smart learning companions (for example, animated intelligent characters/agents) become popular, children are at risk of overestimating and over-trusting AI given their tendency to anthropomorphize such intelligent systems. This CAREER project investigates the development of novel embodied learning technologies that help K-12 students demystify machine learning (ML), an integral aspect of current approaches to AI. The project will provide hands-on and collaborative learning experiences for children to make sense of the inner workings of ML, similar to how they build, act, and experiment in collaboration with friends. The learning experiences will be designed to be accessible to children, regardless of their math and computing background, with special attention to those from historically underrepresented backgrounds in STEM. The project outcomes will advance an AI-driven society by preparing 21st century learners to become critical thinkers about AI, as both consumers and future creators. It will also promote inclusion in next-generation STEM education by addressing AI inequality in life and work.

AI and ML are often presented to children as a black box, focusing on workflows and capabilities (e.g., data training, image and voice recognition) which can lead to inaccurate or oversimplified understandings. Further, many K-12 students lack math and computing backgrounds that are required for understanding abstract Machine Learning (ML) concepts and methods. To address these challenges in understanding abstract ML concepts, this project will explore the design space of 3D and tangible interaction technologies to provide embodied learning experiences that draw upon children's real-life experience of object manipulation, body movement and role-play. Knowledge discovery will be accelerated through a pedagogical agent with curiosity-eliciting prompts to encourage exploratory learning. The learning experiences will be evaluated for impact in supporting knowledge acquisition, self-efficacy and interest in ML with elementary and middle school students. Findings of this project are expected to: (1) deepen knowledge of embodied and exploratory learning in supporting the understanding of abstract and complex STEM concepts through the lens of ML education; and, (2) inform the design of future learning technologies that seamlessly integrate sensorimotor enactment and situated social prompts to make K-12 ML education highly accessible to students with diverse backgrounds and skills.

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.