Nerd Culture and Bias

My father is in the tech business. His company, Raybeam builds the algorithms that help optimize search engine results and target advertisements. He went to Yale for biochemistry, and his college roommate is a Nasa engineer who helped design the Curiosity Rover. My (white) cousin owns an anime shop in Rhode Island, and my (Korean) cousin's Japaneese wife was an employee at Google before joining the team that created the successful fashion app, Polyvore.

This is all to say, I am familiar with the culture of the tech industry, if not the inner workings of technology itself. If my dad weren't so busy at work this week, I would like to interview him about his company culture and gender dynamics.

But it is almost three o'clock on Wednesday, so I only have my own observations.  My dad's company has twenty eight employees, and, as far as I can tell, only four of them are female. They play Magic: The Gathering during their lunch hour and pitched in to buy an arcade console for the office. Three of the women work in communications and public relations, and are recent college graduates. Only one, an older woman, works with software. This pattern basically falls in line with the industry norm as presented by Ms. Kass on Monday. I have met some of my dad's employees, and to be honest, we really get along.

However, I am not surprised that the algorithms that this group of people develop and utilize reflect society's biases, for several reasons.

First, as Ms. Kass mentioned, demographics. White males are going to reflect white maleness in their coding. This is just human nature.

Second, demographics, but for a different reason. What Ms. Kass did not mention during her presentation is that not only are members of the tech industry primarily white and male, but so are technology consumers. As a company that markets to a consumer base of a majority of white males, search algorithms designed to learn from their user-base are going to learn to cater to white males.

Third, business. Changing sample composition means taking more time, which means taking more money. Businesses, in general, are going to take the most expedient route. Many smaller business literally do not have the time or the money to modify their sample composition. The will to change data practices has to come from the consumer, but the actual changes must come from the top. Companies such as Facebook, Google, Apple, and Microsoft, who can afford to take big risks must defy expectations and take the biggest business risk of them all: doing the right thing.