Designing Interactive Visualizations of Neural Pathways in Language-Based AI for Secondary Students to Explore Interpretability of AI and Human-Machine Collaboration

Artificial Intelligence (AI) is transforming numerous industries and catalyzing scientific discoveries and engineering innovations. To prepare to enter an AI-ready workforce, young people must be introduced to core AI concepts and practices early to develop fundamental understandings and productive attitudes. Neural networks, a key approach in AI development, have been introduced to secondary students using various approaches.

However, more work is needed to address the interpretability of neural networks and human-machine collaboration in the development process. This exploratory project will develop and test a digital learning tool for secondary students to learn how to interpret neural networks and collaborate with the algorithm to improve AI systems. The learning tool will allow students to interact with complex concepts visually and dynamically.

It will also leverage students’ knowledge and intuition of natural languages by contextualizing neural networks in natural language processing systems. The project team includes learning experience designers and technology developers from the Concord Consortium, computer scientists from Carnegie Mellon University, educational researchers from North Carolina State University, curriculum specialists and teacher educators from Mississippi State University Center for Cyber Education, and usability and feasibility evaluators from WestEd.

Two middle school teachers from Massachusetts and Mississippi and over 50 eighth grade students will be directly impacted through their participation as co-designers or testers.

This project will investigate the design of learning tools and learning experiences for middle school students to engage with neural pathways and human-machine collaboration in AI development. Using design-based research and participatory design methods, the project will address the research question: What are the characteristics of learning tools that can support middle school students in developing an understanding of neural pathways in language-based AI and competencies in human-machine collaboration in AI development?

The project team will (1) develop and test interactive visualizations of neural pathways for students to investigate neural pathways with unigrams and word embedding as the input layers; (2) iteratively enact and improve the design with five student volunteers and two middle school teachers participating as co-designers and testers; (3) conduct classroom testing with the two co-design teachers in their classrooms (with approximately 50 students). In both lab and classroom tests, project staff will develop instructions and learning activities, facilitate testing sessions, and collect observation, interview, survey, and video/screencast data.

The data will be analyzed qualitatively and quantitatively to inform the revision and refinement of both theory and design. The developed learning tool and exemplary learning activities will be made freely available and contribute to K-12 AI education resources and knowledge base that benefit all students, especially those from demographic groups underrepresented in the computing field, to develop their talent and interest in AI and computer science.

This project is funded by the Discovery Research preK-12 (DRK-12) program, which seeks to significantly enhance the learning and teaching of science, technology, engineering and mathematics (STEM) by preK-12 students and teachers, through research and development of innovative resources, models and tools. Projects in the DRK-12 program build on fundamental research in STEM education and prior research and development efforts that provide theoretical and empirical justification for proposed projects.

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.