Matt Dube, Associate Professor of Data Science, Computer Information Systems (CIS), and Applied Mathematics, and Rocko Graziano, Lecturer in Data Science and CIS, have published “Identification: A Teaching Moment for Privacy and Databases” in ACM’s Digital Education Library EngageCSEdu, a platform for sharing course materials with computer science instructors.
The lesson tackles the critical topic of privacy by exploring identification concepts within databases and association rules within data mining and machine learning. In both contexts, an ethical issue is neglected: the securing of one’s right to privacy. The exercise exploits the notion of conditional functional dependency and the ability in an open data environment of connecting resources that were never meant to be connected. It also highlights the risk of proxy discrimination, where algorithms are trained using datasets which reflect and perpetuate past societal bias.
Their paper will be presented at the EngageCSEdu Technical Symposium in March.