Abstract
An oft-repeated claim among certain social scientists is that we are now in the age of “big data.” In The Black Box Society, Frank Pasquale argues that although this influx of information has benefited societies in many ways, such as through improved health care delivery, it has simultaneously exacerbated particular social problems, including the subprime mortgage crisis of 2009. Throughout the book, Pasquale extends what we know about these problems by arguing that they stem from or are intensified by the improper use of algorithms, rule-based programs that reduce data’s complexity for making automated decisions.
Pasquale develops this argument across the book’s four parts. In the first section, he explains how and why tremendous amounts of information about individuals is beingcollected, sold, and used for multiple purposes—mostly without the awareness or consent of the individuals to whom such data refers. Pasquale highlights how uncertainty, commensuration, opacity, and access to information problematize the application of algorithms to decision-making programs. In the second and third parts of the book, which respectively pertain to algorithms in information and financial systems, Pasquale argues that the corporations and organizations that obtain these data use it for secret algorithms, which are hidden either because they are confidential (e.g., the National Security Agency’s Internet surveillance program) or proprietary (e.g., Wall Street firms’ investing software), that make decisions that affect outcomes ranging from who sees what information and advertisements on the Internet to how large corporations make financial decisions that affect markets and economies, on an international scale, and individuals’ livelihoods, on a local scale. Pasquale develops these arguments with analyses of case studies and primary and secondary materials from legal cases, government and business documents, and news reports.
In the fourth and final section, Pasquale presents several possibilities for how we can fix the algorithmic problems he identifies. One option is to increase regulations on how algorithms are used. Another option is for markets to increase individuals’ access and control over their personal data, which would give individuals a choice between keeping their information private or selling it to others. The best option, Pasquale believes, likely incorporates elements of both.
The Black Box Society is timely, and the rich detail with which Pasquale describes his cases can help us understand why recent antitrust lawsuits have been filed against Google in Europe, events which occurred after the book was published. Importantly, the book is highly accessible and is written in such a way that individuals both with and without knowledge of algorithms can learn much. Those who would benefit most from it, however, include people with interests in organizations, markets, stratification, and computational sciences.