An Empirical Analysis of High School Students' Understanding of A.I. Through Modeling Texts

• In Event: AERA Poster Session 10
  In Poster Session: SIG Poster Session on Data Science and A.I. in Education (San Diego)

Abstract

This study contributes to the growing interest in K-12 Artificial Intelligence (AI) education by examining high school students’ understandings of AI through modeling texts with the XXXX (omitted for blind review), a web-based text-mining and narrative modeling technology. The XXXX curriculum and technology were designed to engage students in text mining in an accessible and engaging way. In this study, we developed a curriculum module that introduced the machine learning workflow of modeling texts with a focus on feature engineering and the role of human insights. This module was implemented at a public high school in the United States. Our qualitative analysis of the interviews, classroom observations, surveys, and assessments revealed that students developed in-depth understandings of how AI works.

Authors

• Cansu Tatar, North Carolina State University

• Michael Miller Yoder, Carnegie Mellon University

• Duncan Culbreth, North Carolina State University

• Kenia Wiedemann

• Jie Chao, The Concord Consortium

• William Finzer, The Concord Consortium

• Shiyan Jiang, North Carolina State University

• Carolyn Rosé, Carnegie Mellon University