Body
Image
Engaging High School Youth in Algorithmic Justice Through Audits of Designed and Everyday Machine Learning Applications
This project investigates how youth engage with algorithm auditing, a method that involves repeatedly querying AI/ML algorithmic systems and observing their output in order to draw conclusions about the system's opaque inner workings and possible external impact. We investigate how youth audit everyday ML applications. More specifically, (1) the feasibility of user-led algorithm audits by youth, (2) the dynamics of collaboration in algorithm audits, and (3) youth understanding of algorithmic justice through auditing.
Research Questions
- What are high school youth experiences and understandings of everyday ML applications?
- How do high school youth design and conduct collaborative audits of ML applications?
- How can high school youth apply audit approaches to applications they encounter in their everyday lives?
Accomplishments to date
- Conducted participatory design sessions with 7 high school youth to design algorithm auditing learning activities.
- Piloted algorithm auditing learning activities during 6 after school workshops with 21 high school first year students.
- Developed a sequence of steps to involve youth in algorithm auditing activities.
Equity
- Participants are majority urban youth of color, a group systematically underserved in STEM
- Investigating algorithmic justice and bias issues
Future work
- Integrate auditing activities into youth design processes of ML applications (summer workshop).
- Co-design auditing activities with teachers (see NSF grant #2342438).
- Integrate and implement auditing activities in formal classrooms (see NSF grant #2342438).
