© 2020 Strange Loop
Increased state use of facial recognition technology threatens to increase the power dynamic between the state and disenfranchised communities. However, what happens when individuals use facial recognition to watch the watchers? OpenOversight is an open-source project that promotes police accountability through public data. Users can sort through photos to identify an officer against whom they would like to complain. A lack of officer identification information has stifled previous attempts to monitor police; in Chicago from 2011 - 2015, 28% of complaints against police were immediately dropped due to a lack of identification. The OpenOversight team has leveraged facial recognition to make the project more effective. For example, we use the AWS tool Rekognition, which has been used by various police departments, to identify with 90% accuracy if police or military officers are present in a photo. This pre-processing has made the volunteer job of photo sorting more efficient. We also use face-api.js to match new photos against existing images in the database. After a discussion of this technical implementation, the talk will close by discussing legal considerations that we encountered, such as department-specific guidelines and compliance with Illinois's Biometric Information Privacy Act. The talk will provide concrete tools and strategies for community responses to state surveillance.