Machines that Judge Us
summary
An art installation gives invisible decision systems audible voices and shows bias in action: three devices run face detection, estimate age and gender, then assemble spoken output with a core.logic dialog system and text‑to‑speech, with vision done on Raspberry Pi + OpenCV and components talking over HTTP local network. The setup runs in a gallery, uses only open source tools to avoid third‑party APIs, provokes surprised and often hostile reactions, and demonstrates how simple rules encode harmful assumptions.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:04Machines that Judge project
- 2:48Software and bias assumptions
- 4:14Machine errors vs human impact
- 6:52Anthony Grimes overview
- 7:34Technical setup and vision
- 9:00Live demo
- 10:02Intent behind judgments
- 12:33Bias in algorithm design
- 13:16Open source and ethics
- 15:27Responsibility in vision tech
- 16:34Closure 2D library
related talks
34mI, Human
24mDesigning for Transparency in Machine Learning
41mDesign for Cognitive Bias
22mNothing for Us, Without Us; Breaking Unconscious Bias in Building Products
47mSystems of Harm
01h 20mARTOUNTABILITY: ACCOUNTABILITY, AI, AND ART
12mHow we infuse our biases in the products...
28mThe Exhibition as a Living Organism