Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Learning About Algorithm Auditing in Five Steps: Scaffolding How High School Youth Can Systematically and Critically Evaluate Machine Learning Applications

Luis Morales-Navarro, Yasmin B. Kafai, Lauren Vogelstein, Evelyn Yu, Danaë Metaxa

TL;DR

This paper tackles how to empower youth to critically evaluate ML-powered systems through algorithm auditing. It introduces a five-step, learner-centered auditing framework derived from expert and user-driven practices. A two-week design-based case study with 14 high school youths auditing peer-designed TikTok filters demonstrates how the steps can be scaffolded in after-school settings, including hypothesis development, input design, testing, data analysis, and reporting. The findings highlight both the benefits of structured auditing for systematic evaluation and the practical challenges of classroom deployment, platform constraints, and collaborative data generation, with implications for scalable integration into K-12 contexts.

Abstract

While there is widespread interest in supporting young people to critically evaluate machine learning-powered systems, there is little research on how we can support them in inquiring about how these systems work and what their limitations and implications may be. Outside of K-12 education, an effective strategy in evaluating black-boxed systems is algorithm auditing-a method for understanding algorithmic systems' opaque inner workings and external impacts from the outside in. In this paper, we review how expert researchers conduct algorithm audits and how end users engage in auditing practices to propose five steps that, when incorporated into learning activities, can support young people in auditing algorithms. We present a case study of a team of teenagers engaging with each step during an out-of-school workshop in which they audited peer-designed generative AI TikTok filters. We discuss the kind of scaffolds we provided to support youth in algorithm auditing and directions and challenges for integrating algorithm auditing into classroom activities. This paper contributes: (a) a conceptualization of five steps to scaffold algorithm auditing learning activities, and (b) examples of how youth engaged with each step during our pilot study.

Learning About Algorithm Auditing in Five Steps: Scaffolding How High School Youth Can Systematically and Critically Evaluate Machine Learning Applications

TL;DR

This paper tackles how to empower youth to critically evaluate ML-powered systems through algorithm auditing. It introduces a five-step, learner-centered auditing framework derived from expert and user-driven practices. A two-week design-based case study with 14 high school youths auditing peer-designed TikTok filters demonstrates how the steps can be scaffolded in after-school settings, including hypothesis development, input design, testing, data analysis, and reporting. The findings highlight both the benefits of structured auditing for systematic evaluation and the practical challenges of classroom deployment, platform constraints, and collaborative data generation, with implications for scalable integration into K-12 contexts.

Abstract

While there is widespread interest in supporting young people to critically evaluate machine learning-powered systems, there is little research on how we can support them in inquiring about how these systems work and what their limitations and implications may be. Outside of K-12 education, an effective strategy in evaluating black-boxed systems is algorithm auditing-a method for understanding algorithmic systems' opaque inner workings and external impacts from the outside in. In this paper, we review how expert researchers conduct algorithm audits and how end users engage in auditing practices to propose five steps that, when incorporated into learning activities, can support young people in auditing algorithms. We present a case study of a team of teenagers engaging with each step during an out-of-school workshop in which they audited peer-designed generative AI TikTok filters. We discuss the kind of scaffolds we provided to support youth in algorithm auditing and directions and challenges for integrating algorithm auditing into classroom activities. This paper contributes: (a) a conceptualization of five steps to scaffold algorithm auditing learning activities, and (b) examples of how youth engaged with each step during our pilot study.

Paper Structure

This paper contains 20 sections, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background
  3. Auditing Algorithmic Systems in Five Steps
  4. Step 1: Developing a Hypothesis
  5. Step 2: Generating a Set of Systematic, Thorough, and Thoughtful Inputs
  6. Step 3 - Running a Test
  7. Step 4 - Analyzing the Data
  8. Step 5 - Reporting the Results
  9. Methods
  10. Context & Design
  11. Data Collection & Analysis
  12. Findings
  13. Step 1: Developing a Hypothesis
  14. Step 2: Generating a Set of Systematic, Thorough, and Thoughtful Inputs
  15. Step 3 - Running Tests
  16. ...and 5 more sections

Figures (4)

  • Figure 1: Ziyi tested the filter with the default images available on Effect House. Figure displays two input and output pairs.
  • Figure 2: Inputs generated by Ziyi and Ishmael on the two-axis organizer.
  • Figure 3: Table with input and output pairs with notes.
  • Figure 4: Input-output pair for the first test conducted by Ziyi and Ishmael.