Book Review: Race After Technology: Abolitionist Tools for the New Jim Code by Ruha Benjamin

Race After Technology: Abolitionist Tools for the New Jim Code by Ruha Benjamin

My rating: 5 of 5 stars

This book builds on the research in Algorithms of Oppression: How Search Engines Reinforce Racism and Dark Matters: On the Surveillance of Blackness, so I definitely recommend reading those two books first. I’m not alone in that, in one of the talks I’ve watched Benjamin give, she explicitly mentions those books as influencing her. I really enjoyed this book, it brought together ideas from my own master’s degree, including the complexity of how technology is used. In one class we specifically discussed the Moses’s bridges in New York (despite this being taught in the Netherlands), which were designed to exclude the poor by preventing buses from crossing the bridge. In this book she discusses this bridge and how it can pull in the very people that were expected to benefit the bridge design (basically a bus full of rich white kids went across after they came back from a trip to Europe, the driver hit the top of the bridge which resulted in 6 people getting seriously injured).

She modernizes these examples by describing how algorithms are created to approximate details about people, such as determining their ethnicity to provide “targeted services.” Due to historical redlining, the practice of creating white people only enclaves in suburbs and portions of the city (a Jim Crow era set of laws), the zip code has become a reliable indicator of ethnicity and race. She gives the example of Diversity, Inc., which creates ethnicity or racial classifications for potentially hiring companies. They will look at the names of people and assess their ethnicity, however due to the history of slavery, many African Americans have white sounding surnames, like Sarah Johnson, to “correctly” identify the ethnicity of Sarah, the company uses her zipcode to assign her race.

Overall, I found a lot of examples in this book very illuminating. Benjamin finds the approach to Design favored in Silicon Valley wanting and excluding, primarily focused on empathizing for making money, which in many cases is empathizing with whiteness. Furthermore, Benjamin argues that empathy can lead skewed results, such as body camera video providing empathy for police officers even when they are killing Black people for crimes which aren’t capital offenses or no crime at all.

As an engineer, I took this book as a warning. That we need to understand how data is impacting those around us. That we need to understand how data that might seem harmless to me, could cause serious harm to someone else. That algorithms that seem to be doing good, could instead be quickly turned into something bad. Facial recognition is a great example. Facebook tags people in photos without consent and this can be exploited by law enforcement. Furthermore, since facial recognition software is so inaccurate, it can misclassify a person as the wrong sex, the wrong person, or in extremely bad past cases, as an animal.

Furthermore, engineers have the responsibility to ensure our work is used to create more equity in the world. Benjamin offers a few different organizations that are working to ensure justice and equity for everyone. Maybe it’s time that software engineers/developers have a responsibility for this the same way a civil engineer must ensure a bridge is safe.

I recommend that anyone that works at a social media company read this. Anyone doing work for algorithms in banks, insurance, hiring, and housing really understand the fact that algorithms aren’t objective. They are as objective as our history. Our history hasn’t been objective nor equitable. We must change that.

View all my reviews

Innovation, Science and Money II

In my last blog I discussed some of the budgetary cuts occurring in the US and how these cuts are going to impact the future of science. I want to spend some time explaining why this is the case. I mentioned something called Path Dependency, what do I mean by this? Well it’s a pretty simple concept, once you start down a policy path your choices are constrained by your previous choices and the results based on those choices.

This type of path dependency can be seen in scientific and technological changes. For example, if a piece of technology has three parts each one can be improved independently. If each one can be changed in one direction, from a 0 to a 1 each change could impact how likely a specific technology would be selected by consumers. Each change could lead to a local optimal, and could prevent the technology from becoming a global optimal. Additionally, these changes over time, with further research, could lead to radical different technologies. This happening from changing a single feature from on or off. Basically, it’s an evolutionary process.

Policy works the same way. There’s a paper written by Mustar et al (2008) that discusses the policy choices in France and the UK. The objective of the paper was to investigate the impact of policy choices on the creating of academic spin-offs. Some of the results lead to additional technology incubators in the UK and in France. However, the number of academic spin-offs in France actually decreased, however in the UK they increased significantly.

These differences came about because of previous policy choices. For example, France has laws related to civil servants and starting a new company. In France all professors are considered civil servants, so there is a history of professors not starting companies. There’s a lack of culture for entrepreneurship in France for increasing the number of academic spin-offs.

This is what I meant by path dependencies. Decreasing the amount of money going into meaningful academic research will have an impact in other ways. In the US there has been an increased push for increasing the number of companies being started. Scientific research can be turned into new companies through academic spin-offs. Decreasing the funding at two of the biggest funding agencies will decrease the number of academic spin-offs.

Mustar et al 2008