More Than a Glitch: Confronting Race, Gender, and Ability Bias in Tech — A Book Review

This book is about the many ways that technochauvinism has an adverse effect on the creation of technology, and it offers ways to intervene. Unfortunately, there isn’t a magic switch that we can flip to make technology work for everyone, everywhere. Eliminating racism and ableism and gender bias from technological systems is a complex problem. Race as a single subject is complicated. So is technology, and so are gender and disability. The intersection of all of these concepts is hard to understand. So much so that it needs the space of a book to untangle the threads. — Meredith Broussard, More Than a Glitch: Confronting Race, Gender, and Ability Bias in Tech (pg. 7)

While we are being bombarded with articles on ChatGPT and other artificial intelligence based solutions to society’s problems, Meredith Broussard’s new book More than a Glitch is an excellent antidote to the hyperbole. Her examination of the many missteps in computing covers decades of computing solutions and the thinking behind them.

Computers add numbers very fast. How they add numbers is programming, based on algorithms. What numbers they add together is a dataset, based on what information the programmers consider important. Both the compiling of datasets and the writing of computer programs is based on human cognition which is notoriously limited and biased even for the best of us. Repeatedly throughout the past several decades programs designed to help have caused harm. And protected classes and the most vulnerable in society are usually the ones harmed.

Broussard’s background in data journalism and the field known as artificial intelligence ethics gives her a broad perspective on the flawed computing solutions that deepen the already existing biases and prejudices in society. She stays away from the legal questions of blame and fault in these failures. To combat technochauvinism, which Broussard describes as “a kind of bias that considers computational solutions to be superior to all other solutions,” she asks for readers to be very critical of when a computing solution is necessary and how it should be deployed.

From motion sensing soap dispensers that can’t see brown skin to facial recognition software that consistently misidentifies and misgenders persons. From standards in web design that reject those with sensory limitations to interfaces that deny access to those with physical restrictions. There are so many instances of failure in using computing technology to solve problems. And too often the predictive algorithms are based on flawed statistics and biased data. Affordable mortgages and appropriate health care are denied to persons of color, and future guilt is determined by predicting who will commit crimes before they happen.

This book is a well written, concise look at some of the more serious cracks in the foundations of computing. The fallout from these technologies affects everyone, even if they never touch a computer in their lifetime. (LJ)