Emotional Machines: AI Meets DEI

Purpose & objective

With the explosion of artificial intelligence tools over the past year, it has become increasingly important for students to understand how these systems work. In this interactive session, participants will be introduced to a code-free exploration of machine learning and bias. In teams, participants will use the Machine Learning for Kids online tool to train a machine learning model to classify faces based on emotion. The goal for this session is to provide a code-free way to start conversations with your students about artificial intelligence, how it works, and how bias can be intentionally and unintentionally built into AI. This session is designed for teachers from elementary to high school and at any point along the computer science skill spectrum…no prior knowledge needed!

Outline

In this interactive workshop, participant teams will train a machine learning model to classify human faces based on the 4 Zones of Emotional Regulation: green, yellow, blue, and red. These zones are commonly used in elementary schools to teach students about SEL.

Introduction (10 minutes): Overview of AI Basics and modeling of technology

Data Collection (15 minutes): Teams will collect image data (i.e., faces), train, and test their model using Machine Learning for Kids..and add more images if necessary!

Beta Test (15 minutes): Teams will rotate to test each other's facial recognition model. If the model is not successful, they can add more images.

Debrief (10 minutes): Teams will return to their computer and see how their AI model has changed. Group discussion of bias through data collection.

Closing/Q&A (10 minutes)

Supporting research

Saving face: Investigating the ethical concerns of facial recognition auditing
ID Raji, T Gebru, M Mitchell, J Buolamwini, J Lee… - Proceedings of the AAAI/ACM Conference on AI, Ethics …, 2020 (https://dl.acm.org/doi/abs/10.1145/3375627.3375820)

Code.org Materials on AI impact:
https://code.org/ai/pl/101