97 Things About Ethics Everyone in Data Science Should Know

Chapter 21. Algorithmic Bias: Are You a Bystander or an Upstander?

Jitendra Mudhol and Heidi Livingston Eisips

Founder and CEO, CollaMeta Adjunct Faculty, San Jose State University

Automated Decision Systems (ADS) are used in many human endeavors, from ad targeting and credit scoring to child welfare and criminal justice. ADS impact us daily, but we seem oblivious to the pervasive influence of algorithms on our lives: up to 87% of Americans are identifiable from their zip code, birthday, and gender.¹

A well-designed algorithm can bring positive socioeconomic change; for instance, machine learning is at the heart of many medical imaging and drug discovery innovations.² Still, ADS may be amplifying bias on an unprecedented scale, while giving it the garb of scientific objectivity. In Automating Inequality, Virginia Eubanks “exposes how US institutions, from law enforcement to health care to social services, increasingly punish people—especially people of color—for being poor.”³

We face a choice: either act to address algorithmic bias or turn a blind eye. Psychologists call the latter bystander apathy.

Understanding Bystanderism

Research in the late 1960s revealed that the greater the number of people present at an emergency, the lower the likelihood of receiving help.⁴ Scott Lilienfeld ...