Investigating Youths' Everyday Understanding of Machine Learning Applications: a Knowledge-in-Pieces Perspective

TL;DR

This paper explores how teenagers understand machine learning in everyday contexts, revealing partial knowledge about data training and pattern recognition, which can inform educational strategies for AI literacy.

Contribution

It applies a knowledge-in-pieces framework to analyze youths' informal understanding of ML, highlighting productive knowledge aspects for education.

Findings

Teens understand ML learns from data

They recognize pattern-based outputs in ML applications

Some understanding of ML's data-driven nature exists

Abstract

Despite recent calls for including artificial intelligence (AI) literacy in K-12 education, not enough attention has been paid to studying youths' everyday knowledge about machine learning (ML). Most research has examined how youths attribute intelligence to AI/ML systems. Other studies have centered on youths' theories and hypotheses about ML highlighting their misconceptions and how these may hinder learning. However, research on conceptual change shows that youths may not have coherent theories about scientific phenomena and instead have knowledge pieces that can be productive for formal learning. We investigate teens' everyday understanding of ML through a knowledge-in-pieces perspective. Our analyses reveal that youths showed some understanding that ML applications learn from training data and that applications recognize patterns in input data and depending on these provide different outputs. We discuss how these findings expand our knowledge base and implications for the design of tools and activities to introduce youths to ML.