Technically Wrong: Sexist Apps, Biased Algorithms, and Other Threats of Toxic Tech by Sara Wachter-Boettcher

 

Fun fact: I applied for an ARC when this was first coming out and did not get a copy. No, I'm not salty about it (she said, while casually presenting like a human-shaped lump of salt). When this book went on sale, you can bet that I snapped it right up, hyped as all get out to read it, because as a feminist who works in the vast technology industry, this was relevant to all my interests. And it's good. Even though you've probably seen most of the case examples Wchter-Boettcher brings up when they made the news (such as the lens-focusing cameras that didn't work on Black people, or Twitter's failure to stop the mass-harassment of many women online, including Lindy West). And even though the "beware the evils of technology horse" is the latest and greatest dead horse that the news media loves to beat. But it's so useful to have all that information organized into one place, for easy reference, with commentary on how it happened and why it matters.

I'm not sure how appealing this book will be to people who aren't interested in social sciences and social justice and/or who aren't in tech or use it actively for work. But since most of us engage with some sort of app or website on the daily, I think it's important for everyone to understand how technology can fail us. One of the prevailing themes in this book is how a lack of unbiased and undemocratic (small d, don't @ me) information can lead to unconscious bias that reveals itself in AI systems. The author gave the example of a risk-assessment website that predicted what the recidivism of a criminal might be, but it did not take into account racial biases, such as the fact that Black individuals tend to get arrested more regardless of whether they are guilty or not, making one of the value points-- associating with people who have prior arrests-- questionable at BEST. Another is that a lack of diversity can contribute to erasure, or-- worse-- systemic microaggressions that can make a site significantly less safe or useful (examples: binary gender options upon signing up with a website or period-tracking apps that assume their users are straight women who want to get pregnant).

This book was published in 2017 so obviously it's aged. There are some comments about Twitter and its failure to sell the company based on its bad track record, for example, that have aged like a used diaper left out in a hot sun. (I can only imagine what the author thinks of some of the latest site "updates.") For the most part, though, I think a lot of the book is still relevant. Especially with more and more companies leaning more heavily on the use of algorithms and AI. Technology itself is not inherently harmful, but wielded in the wrong hands and used with carelessness, it very easily can be.

4.5 out of 5 stars