Weapons of Math Destruction Summary
Weapons of Math Destruction is a book by Cathy O’Neil that examines the dangers of using “Big Data” to make decisions that can alter the course of people’s lives.
- “Weapons of Math Destruction” (or WMDs) are Big Data models that use algorithms to make decisions. WMDs ignore nuance and often rely on flawed data.
- Many WMDs are opaque to the people whose lives they affect, yet they are everywhere: in finance, college admissions, advertising, police departments, the job market, politics, and social media.
- O’Neil concludes that the use of WMDs is biased, careless, and ultimately unethical.
Last Updated on August 14, 2019, by eNotes Editorial. Word Count: 1239
Cathy O’Neil’s nonfiction book Weapons of Math Destruction exposes how data collection is being used to undermine individuals’ right to privacy and the democratic society in which we live. Weaving her personal experiences with research, O’Neil traces the development of big data and its far-reaching impacts on the average citizen.
Writing an essay?
Get a custom outline
Our Essay Lab can help you tackle any essay assignment within seconds, whether you’re studying Macbeth or the American Revolution. Try it today!
In the introduction of the book, O’Neil describes her lifelong love for numbers, explaining how this led to a career in the field of mathematics. After the housing crisis and economic downturn of 2008, however, O’Neil recognized that mathematical equations were not infallible in measuring human behavior. She explains how, after 2010, the public accepted the growing trend of corporations and public services using algorithms to dictate much of their lives.
She cautions, however, that the ubiquity in these algorithms does not mean they are harmless. To illustrate this, she discusses the educational reforms in Washington, DC, under Michelle Rhee, whose teacher evaluation system IMPACT was meant to identify the lowest-performing teachers in the school district using math. Instead of solving the district’s student achievement problem, the system fired good teachers based on a number that was calculated using undisclosed methods and unreliable data. The ability of mysteriously guarded algorithms to upend an individual’s life without much proof is what inspires O’Neil to write about this subject.
In “Bomb Parts: What Is a Model?” O’Neil discusses how statistical analysis became integral to the sport of professional baseball. She suggests that baseball’s use of data to analyze players’ performance serves as a contrast to the destructive power of weapons of math destruction (WMDs) because it is transparent. In the rest of the chapter, O’Neil explains how models are created with goals and ideology in mind, which means they are just as biased as the people who create them. Because of this, O’Neil asserts that it is dishonest and immoral to keep the details of the model’s creators and purpose from the people whose lives the model will impact. O’Neil discusses recidivism models in prisons as an example of how models can have detrimental consequences, even if they are unintended.
In the next chapter, “Shell Shocked: My Journey of Disillusionment,” O’Neil goes into detail about how she arrived at her suspicions about Big Data. In her time working in hedge funds, O’Neil witnessed the havoc that models and algorithms could wreak on the economy during the 2008 financial crisis, not to mention the human suffering this caused. After leaving Wall Street, O’Neill worked in e-commerce, calculating the likelihood of website visitors to buy products. Even still, O’Neil grew increasingly uncomfortable with the way numbers were being used to manipulate, quantify, and control people—leading her to use her mathematical prowess to lead the charge in investigating Big Data.
In the next chapter, “Arms Race: Going to College,” O’Neil explains how US News undertook the task of ranking the nation’s colleges and universities based on a complex system of mathematical equations. O’Neil explains how the proxies by which the newspaper evaluated each school led to increased enrollments at some schools like Texas Christian University, but that the rankings inadvertently led to the rapid inflation of college tuition. Furthermore, the rankings tightened college admissions criteria, which in turn caused anxious students and their parents to spend exorbitant time and money on ensuring admittance—sometimes even resorting to cheating.
The next chapter, “Propaganda Machine,” explores the online advertising industry and how it uses consumer clicks and online activity to manipulate viewers and drive traffic. O’Neil explains how for-profit colleges prey on low-income and immigrant students with their advertisements, seizing on their vulnerabilities in order to generate profits. To do so, these colleges use demographics acquired via Google or Facebook in order to specifically target the populations most likely to be swayed by their ads. O’Neil uses the lead generation predatory model of the for-profit college to illustrate how Big Data is using our personal weaknesses against us in order to make money.
Chapter 5 tackles how Big Data impacts the criminal justice system. The goal of many models used to analyze crime data is to prevent crime. As a result, many of these models determine the likelihood of an individual to commit a crime in the future, which in turn causes law enforcement to surveil targets who are often already subjected to unfair treatment due to their circumstances. O’Neil explains how the goal of improving community trust in police would be a better one, yet she notes the difficulty in quantifying trust.
In “Eligible to Serve,” O’Neil examines how data impacts hiring practices. She explains how personality tests, credit checks, and other models are used to exclude applicants based on faulty biases; for example, that having a mental health diagnosis means one will be an unstable employee. Because of this, employers have been able to blame their prejudicial hiring practices on data.
Next, O’Neil explains how data analysis has changed the way businesses schedule their employees’ shifts. Analyzing ever-changing data about customer traffic, weather patterns, and special events, these models have distilled scheduling into a science intended to minimize labor costs. Because of this, employees in today’s service industries are often at the mercy of mercurial schedules that interrupt their lives. Furthermore, she explains how value-added systems of evaluating employee performance, such as those used in education, focus less on fairness and accuracy and more on efficiency, which often leaves destroyed lives in its wake.
In “Collateral Damage: Landing Credit,” O’Neil explains how FICO credit scores are calculated. In contrast to transparent FICO scores, e-scores are used to calculate the likelihood of default using online behavior and demographics. O’Neil explains how data scientists create models that lump people into tribes instead of judging them as individuals, which can affect everything from loan approval to job promotion. Because algorithms rely on data portfolios that are due with errors, an individual can be unfairly scored.
Chapter 9 tackles the biases of the insurance industry. After explaining insurance’s origins in actuarial science, O’Neil explains how data has allowed insurance companies to classify policy holders into small, unknown groups in order to price gauge based on arbitrary criteria. She addresses how this disproportionately impacts the poor, since credit scores are one of the most popular proxies used to judge policy holders. In addition, behavioral and health data can be used to discriminate based on any number of factors. O’Neil suggests that health data, for instance, could eventually be used to sift through job applicants.
In the final chapter of the book, O’Neil looks at social media algorithms and their impact on civic life. She describes how Facebook’s experimental algorithm tweaks were used to increase voter turnout, demonstrating social media’s power to directly influence political activities. O’Neil then explains how politicians employ Big Data to analyze voter demographics and predict voting activity, allowing them to specifically sway voters who are on the fence. O’Neill describes how microtargeting allows political analysts to anonymously distribute different information to different voters. This microtargeting prevents the public from banding together, which she says undermines the purpose of democracy.
In the conclusion of the book, O’Neil presents several examples of successful models that have had a positive impact before urging readers to recognize that WMDs need stricter oversight and regulations.