ABCDEFGHIJKLMNOPQRSTUVWXYZ
1

BERKELEY HAAS CENTER FOR EQUITY, GENDER & LEADERSHIP
Bias in Artificial Intelligence: Examples Tracker

Purpose:

This tracker outlines examples of documented bias in artificial intelligence, particularly machine learning, systems. It is meant to support business leaders and other stakeholders in identifying examples relevant to their context to understand where / how bias can enter in AI systems, as well as consequences. It is an accompanying resource to our playbook for business leaders on Mitigating Bias in Artificial Intelligence: https://haas.berkeley.edu/equity/industry/playbooks/mitigating-bias-in-ai/.

How to Use:

Tab 2 includes the examples with descriptions for each and links to further information. There are several categories to enable searching and identifying examples, including identity, industry and type of resource (e.g., peer reviewed article, investigative journalism, etc.). These are also listed in tab 3, while definitions and terms are outlined in tab 4.

How We Define "Biased AI":

In our playbook we refer to AI systems that result in either (1) incorrect outputs / predictions for certain populations and/or (2) discriminatory output / predictions for certain populations as "biased AI". Biased AI systems can unfairly allocate opportunities, resources or information; infringe on civil liberties; pose a detriment to the safety of individuals; fail to provide the same quality of service to some people as others; and negatively impact a person’s wellbeing such as by being derogatory or offensive. Biased AI can negatively impact individuals, society and business.

This tracker focuses on examples pertaining to bias in specific AI systems -- primarily machine learning. However, it also includes examples of bias seen in two widely used image datasets, on the basis of research showing that this bias is evident in the outputs of algorithms trained on these images.

How this Examples Tracker was Developed:

This Compendium was developed by the Center for Equity, Gender & Leadership (EGAL) at the UC Berkeley's Haas School of Business. Team members that developed and informed the Examples Tracker include Ishita Rustagi and Genevieve Smith.

If You Have an Example not in the Tracker Please Share It!

We have created this survey tool for folks to share examples with us that we can review for consideration to add to the tracker: https://docs.google.com/forms/d/e/1FAIpQLSe1BPKbzsARSDPGmGkoXQ5F3qzxrGIkedK3bNLQbT5PaMAUWw/viewform

With any questions please reach out to EGAL@berkeley.edu.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100