AI-Augmented Tangible MakeCode (AI-TMC)

Adaptive Guidance for Scaffolding Collaborative Programming ​


​Jin Yu, River Pease, Sriram Manikandan, Jing Xie, HyunJoo Oh 
Submitted to the ACM Interaction Design and Children Conference (IDC ’26) · Under review 
[PDF available upon request]

Overview
AI-TMC is a web-based extension of Tangible-MakeCode that adds artifact-aware, choice-based AI scaffolding to support collaborative, open-ended programming. Learners upload a photo of their tangible block program, generate MakeCode-compatible code, and use three guided entry points to extend features, understand logic, and realign intent through iterative micro:bit testing.

Key Features
  • Three AI Helper entry points (feature extension, code exploration, intent–code alignment).
  • Artifact-aware explanations grounded in the current tangible program. 
  • MakeCode-compatible output (copy/download) for immediate micro:bit testing. 
  • Follow-up questions supported during development (not only button presses). 
Who Is It For? 
  • Middle-school learners (ages 12–14) doing collaborative, open-ended programming.​
​
Research Highlights 
​Iteration loop in practice:
Repeated cycles of assembling blocks, testing on micro:bit, and returning for targeted AI guidance. ​
Collaborative building:
Co-developing program logic by manipulating shared tangible blocks, making ideas discussable and negotiable at the table.
​Artifact-aware AI help:
Suggestions based on the current tangible program.
Teams get next steps that match what they’ve built.
Intent–code alignment moment:
When “our code isn’t expressing our idea yet,” the AI helped learners articulate intent and translate it into concrete next steps for micro:bit testing without taking over authorship. 
Try It
  • Web: Link  (first visit may be slow due to server start-up)
  • ​Tangible Blocks: Link