Read the original document by opening this link in a new tab.
Table of Contents
I. INTRODUCTION
II. BIAS AND DISCRIMINATION
III. MEASURING BIASES
IV. ATTESTING AND ADDRESSING DISCRIMINATION
V. LEGAL PERSPECTIVE
Summary
The document discusses bias and discrimination in AI systems, highlighting the relationship between the two concepts. It explores various causes of bias in AI, such as bias in modeling, training, and usage. The paper also delves into different approaches to measuring bias, including procedural and relational methods. Furthermore, it examines the challenges in addressing discrimination in algorithmic outputs. The authors argue for a cross-disciplinary approach involving legal, social, and ethical considerations to define the relationship between biases and discrimination in AI.