Cathy O'Neil used to work at a hedge fund, in risk analysis, and as a data scientist, in all of which occupations she used Big Data algorithms constantly. She appreciated the convenience of being able to sort huge amounts of complex data in seconds but increasingly began to question whether the models she used, or mathematical models in general, were actually effective.

This led her to research and write Weapons of Math Destruction. O'Neil concludes that while there are legitimate uses for statistical modeling, many of these models produce inaccurate results which cause harm to people and reinforce structural inequality in society. She refers to these models as WMDs, weapons of math destruction. A WMD has three principal characteristics:

1. It is opaque or invisible. This means that the people whose data is analyzed do not know how the model works or perhaps even that it exists at all.

2. It is unfair, damaging the lives of those whose data is being analyzed.

3. It is scalable, having the capacity to grow exponentially.

O'Neil points out that in the upper echelons of society, people are treated as individuals. Elite schools and employers conduct face-to-face interviews and take individual circumstances into account. Poor people, however, are processed en masse using mathematical models.

Therefore, O'Neil's thesis is that Big Data's WMDs, which can turn lives upside down by refusing credit, employment, or education, are not only inherently flawed and unfair, but biased towards the advantaged.