Imagine that you are deciding whether to release a person on bail, grant a consumer a loan, or hire a job candidate. Now imagine your method of making this decision involves using data to algorithmically predict how people will behave—who will skip bail, default on the loan, or be a good employee. How will you know if the way you determine outcomes is fair?
In recent years, computer scientists and others have done a lot to try to answer this question. The flourishing literature on “algorithmic fairness” offers dozens of possibilities, such as testing whether your algorithm predicts equally well for different people, comparing outcomes by race and sex, and assessing how often predictions are incorrect.
Think of a dangerous job. One where workers experience daily risk and suffering. Where the accidental burn, cut, and blood is to be expected— maybe even mundane. Perhaps what comes to mind is a firefighter barreling into a fiery building, a meat packing plant worker who trims sides of beef, or a police officer in a foot chase with an armed suspect.
What likely did not come to mind are the folks who whipped up the plate of palak paneer you dined on last Saturday or baked the croissant you nibbled this morning: chefs, cooks, and other restaurant kitchen workers. But, data from the U.S. Bureau of Labor Statistics finds that workers at dining establishments—about 2.3 million people in the United States— experience comparatively more pain and injury than those of other professions, including the stereotypically dangerous ones I mentioned earlier (see table below).
Work in Progress is a project of the American Sociological Association's Sections on Organizations, Occupations, and Work, Economic Sociology, Labor and Labor Movements, and Inequality, Poverty, and Mobility