In this 60-minute kit, developers and workforce learners build intuition for how AI-generated code can be trusted and audited. In Activity 1, they use a SwiftUI app in Xcode to fit a line to data by hand. They learn the vocabulary of loss, weights and prediction, and see how an outlier introduces algorithmic bias. In Activity 2, they follow the path from linear regression to recurrent networks to transformers, so they can keep human agency over the logic in the age of agentic coding.
Attribution: This Kit was created by: Chi-Ken Lu, Rutgers University–Newark
Usage rights: This Teaching Kit was designed to help educators and learners design, deploy, and use AI with a human-centered approach. It was created in collaboration with a network of higher education and nonprofit organizations and has been made available under the Creative Commons CC0 1.0 Universal license.
Attach up to 5 files which will be available for other members to download.