In Computer Science 2 (CS2) last week, students watched the Documentary Coded Bias from PBS. The documentary follows Joy Buolamwini’s (MIT Media Lab) investigation into algorithmic bias in facial recognition, revealing how artificial intelligence can perpetuate discrimination and highlighting women’s efforts to protect civil rights. After watching the documentary, students discussed how machine learning models are used to create individualized social media feeds and the potential harms that this may pose. One student described what they took away from the documentary as understanding how social media feeds can influence behavior, like voting, to some demographic groups but not others. Watching this documentary was a lead into students doing work with machine learning and AI.
CS2 students are currently engaging in exploring machine learning that asks them to categorize data in order to make predictions. To begin this process, the students completed an activity in which they were role playing zookeepers who have a group of monkeys, some of whom bite and some of whom do not. Using current data, they need to develop rules to then predict whether new, incoming monkeys were going to bite or not. They did this without any computer coding- it was a manual exercise. By using predefined rules and physical sorts, they mimicked the behind the scenes work computer algorithms use to sort data. This visual activity helps to demystify computer processes. From that initial activity, students then learned how to use training, testing and validation data to build the same sort of predictive model for all types of existing data. Next up, students will be taking data from Portugal to predict the likelihood of a student graduating from high school based on demographic data. Students will consider the ethical concerns of this model, and ways that it could help or harm students from Portugal.
