Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy Cathy O'Neil

The head of the computer department at Regis University, where I teach, recommended this book to all faculty. The author, a math nerd, who started her career as an academician teaching at Columbia and the switching to finance, worked as a quant for D. E. Shaw, a hedge fund company. The Great Recession caused her epiphany, when she realized the damage that numbers can inflict. Although numbers present themselves a objective and unbiased, she realized that numbers and the models, the assumptions behind the numbers, "were based on choices made by fallible human beings . . many of these models encoded human prejudice, misunderstanding, and bias into the software systems that increasingly managed your lives." (p. 3)

Through out the book, O'Neil gives examples of how numbers, models, algorithms, and bid data offer false conclusions. Because the numbers and the assumptions behind them usually can be understood only by the individuals who construct them, people who apply them to actual people, do not understand the consequences of the logic. These "weapons of math destruction (WMDs)" have permeated our society.  In education these numbers and assumptions evaluate students and teachers. In the judicial system, they determine convictions and prison sentences. In finance systems, they caused the crash of 2007, stop and frisk searches in large urban centers, U.S. News and World Report college rankings. President Obama unsuccessfully attempted to correct the U.S. News and World Report college rankings by basing rankings on "graduation rates, class size, alumni employment and income, and other metrics" (p. 67). With computerized health and credit data, zip code information, and Facebook friend information, the numbers, algorithms can determine employ ability,  innocence or guilt in the justice system, and by scanning online data, the amount of ideas a person generates, which determines employment desirability. What website a person accesses might affect their FICO or credit score. Call center personnel data found that people who engaged in job-related chatter performed better at their jobs than those who did not.

With all the scores and assessments of individuals, based on data, a dossier of who we are, what we have done, and what we have can contain correct or incorrect information. Will our job, health, credit, education, insurance, safety, news, and social prospects result from who we really are or what incorrect online information attributes to us.