Centering Maker Voices and Ideas with AI-powered Support

• In Event: Empowerment-Focused Computing in Constructionist Learning Environments

Abstract

Constructionist environments center makers’ voices, ideas, and interests. They provide fertile ground for digital empowerment, encouraging makers to connect computing to real-world problems, and recognize their ability to impact their lives and communities (Tissenbaum et al., 2021). However, successfully integrating empowerment into real-world learning settings requires extensive support. We believe Artificial Intelligence (AI) tools may offer new possibilities to provide dynamic support in complex constructionist settings. In this poster, we describe the design and use of a new AI tool to support maker reflection, help-seeking, collaboration, and facilitator-provided scaffolding in constructionist making.
As part of the [Blinded] project we developed the Make-in-Progress (MiP) carousel, a tool that allows makers to share and view images and videos of their in-progress constructions with other makers that may not be co-located. Prior analysis suggests that sharing and viewing media of projects allow learners to showcase their evolving expertise, builds a sense of community, and invites makers to imagine new kinds of computing projects relevant to their lives (Blinded). Here we extend this work by exploring the implications of adding AI generated suggestions for students and facilitators to the MiP.
During a two-week summer workshop with middle-school-aged makers, we connected a private instance of Google's Gemini to the live database of student posts to create a recommendation tool that would suggest project additions and encourage collaboration. For example, analyzing media students have uploaded to MiP the system might suggest: “Jenny’s app could be enhanced by incorporating a MapScreen component (similar to Andre's) that displays local resources (food banks, housing services, etc.) as interactive pins”. Likewise, evaluating recent posts by a student, the system might suggest the maker showcase new emerging skills such as “if/then logic” or values such as “equity and justice.” After being reviewed by the project team, these AI-generated suggestions can be shared directly with students or with facilitators to supplement their awareness and knowledge of the makerspace.
In this implementation we found the suggestion tool provided meaningful contributions to facilitator-driven support. For example, the AI tool encouraged useful and unique additions to student projects and suggested pairings that matched those recognized by makerspace facilitators. However, some suggestions (such as student pairs reluctant to work together) lacked contextual awareness highlighting the importance of facilitator expertise. In our design, a review screen enabled the project team to review AI suggestions before sharing them with students, filtering out irrelevant suggestions. These observations reaffirm the importance of using AI as a support tool for knowledgeable educators, rather than a replacement.
Our findings suggest new opportunities and cautions for integrating AI into support systems in constructionist spaces. By considering AI suggestions, facilitators can offer additional support as makers take on computing projects that matter to them and build peer connections. We see potential for AI integration that amplifies what young makers can do without replacing their voice, centering empowerment.

Authors

• Blake Danzig, Teachers College, Columbia University

• Emmy Semprun, Teachers College, Columbia University

• J K Voorhis, Teachers College, Columbia University

• Nathan Holbert, Teachers College, Columbia University