Fairness and Abstraction in Sociotechnical Systems

Focus: Methods or Design

Source: FAT 2019

Readability: Expert

Type: PDF Article

Open Source: No

External URL: https://dl.acm.org/doi/10.1145/3287560.3287598

Keywords: N/A

Learn Tags: Design/Methods Ethics Framework AI and Machine Learning Research Centre

Summary: An article that outlines five traps that fair-ML work can fall into — framing, portability, formalism, ripple effect and solutionism — even though it tries to be more context-aware than traditional data science.