Weapons of Math Destruction
How Big Data Increases Inequality and Threatens Democracy
Cathy O'Neil · 2016 · Technology
13 min · 13 chaptersNarrated in your voice
What it's about
Cathy O'Neil argues that many of the algorithms now used to sort, score, and judge people are not neutral tools but opaque, unaccountable systems that punish the poor and reward the already advantaged. Drawing on her own career as a Wall Street quant, she shows how models used in education, criminal justice, hiring, lending, and advertising can encode bias, scale it nationally, and hide it behind a false claim of mathematical objectivity.
Read a little every day
15 min a day
1 days · done Sat, Oct 31
Read together
Start a circle and share the code — everyone sees everyone's honest place in the book. Accountability, not leaderboards.
Who it's for
- A data scientist or engineer who builds scoring or ranking systems and wants to understand their social stakes - A policymaker or journalist trying to evaluate claims that an algorithm is 'fair' or 'objective' - A general reader affected by credit scores, hiring software, or predictive policing who wants to understand why
What’s inside
- 1The Crash as a Warning1 min
The 2008 financial crisis shows that mathematical models can cause catastrophic, society-wide harm while being treated as too complex and too authoritative to question.
- 2What Makes a Model a Weapon1 min
A statistical model becomes destructive specifically when it combines opacity, massive scale, and a damaging feedback loop, not simply because it uses data.
- 3Grading the Teachers1 min
The Washington D.C. teacher value-added scoring system punished good teachers with statistically meaningless numbers dressed up as objective measurement.
- 4Rankings That Rank Themselves1 min
The U.S. News college rankings did not just measure quality, they redefined it, forcing schools to compete on metrics that often had little to do with education.
- 5Preying on the Vulnerable1 min
Companies used data-driven targeting to find and exploit the most anxious and financially desperate people, turning personal data into a tool for predatory marketing.
- 6Policing Its Own Prophecy1 min
Predictive policing and recidivism scoring do not predict crime, they predict and then reinforce where police already look and whom the system already punishes.
- 7Screened Out Before You Apply1 min
Automated hiring tools filter out qualified candidates using hidden criteria that can amount to illegal discrimination without anyone intending it.
- 8Scheduled Into Instability1 min
Workforce scheduling software optimizes for corporate efficiency at the direct expense of workers' health, income stability, and family life.
- 9The New Redlining1 min
Modern credit and insurance scoring uses proxies correlated with race and class to recreate discriminatory pricing that fair-lending laws were designed to prevent.
- 10Voting in the Dark1 min
Political campaigns' use of microtargeted data undermines the shared public information that democratic debate depends on.
About Cathy O'Neil
Cathy O'Neil holds a PhD in mathematics from Harvard and worked as a quantitative analyst at the hedge fund D.E. Shaw before moving into data science and joining Occupy Wall Street. She writes the mathbabe blog and founded an algorithmic auditing company, giving her both technical fluency and firsthand experience of how financial models fail.
