Stakeholder-Driven Contextual Evaluation of AI Bias: Novel Conceptions, Methods, Tools

Research Overview

This research explores translanguaging in AI-assisted classrooms.

Objectives

  • Understand how students switch languages while learning.
  • Build AI tools that adapt to language diversity.
  • Enable equitable education through tech.

Image Example

TRAIL Lab Logo

Dataset Sample

{
  "student_id": "abc123",
  "language": "Spanglish",
  "utterance": "Yo pienso que the Earth orbits el sol."
}

Inline code like const x = 5 is also supported.

Results Table

Metric Value
Accuracy 93.5%
Precision 91.2%
Recall 92.7%

External Resources

Future Work

  1. Expand model support for additional languages.
  2. Partner with more bilingual schools.
  3. Publish findings in academic venues.

Generated on: June 19, 2025

2019

1 publication

  1. Using Participatory Design to Facilitate In-service Teacher Learning in Computational Thinking

    Shamya Karumbaiah, Sugat Dabholkar, Jooeun Shim, Susan Yoon, Betty Chandy, Andy Ye

    Proceedings of the 13th International Conference on Computer Supported Collaborative Learning (CSCL)

    View paper

2018

1 publication

  1. Is Student Frustration in Learning Games More Associated with Game Mechanics or Conceptual Understanding?

    Shamya Karumbaiah, Seyedahmad Rahimi, Ryan Baker, Valerie Shute, Sidney Dmello

    Proceedings of the 13th International Conference of the Learning Sciences (ICLS)

    View paper

People

TRAIL Lab

The Responsible AI for Learning Lab — asking whether AI belongs in classrooms, not just how.

University of Wisconsin–Madison · Educational Psychology

Contact

1025 W Johnson St

Madison, WI 53706

shamya.karumbaiah@wisc.edu

View on Google Maps →

Feedback or accessibility issues? shamya.karumbaiah@wisc.edu