Google Street View Could Worsen Political and Racial Divides Between Neighborhoods

1_4_Google Street View
A Google Street View vehicle collects imagery for Google Maps while driving down a street in Calais, northern France, on July 29, 2015. Philippe Huguen/AFP/Getty Images

Google Street View is a helpful tool to help navigate your neighborhood, but it may also shed light on much more unexpected things, such as who your community voted for or even your neighbor's race.

More specifically, the types of vehicles—sedans or trucks—captured by the technology may provide this information, according to research conducted at Stanford University.

Researchers used data from 50 million Street View Images from 200 cities nationwide. In total, they came across 22 million vehicles. That's about 8 percent of all vehicles in the United States. They then used two algorithms to identify the makes, models, and years of the automobiles.

Related: Tour the International Space Station With Google Street View

Timnit Gebru, a postdoctoral researcher at Microsoft Research, New York City and her colleagues, then linked the information with other data, including presidential election voting records. Using artificial intelligence algorithms, they determined voting patterns among various precincts where the cars were. According to their findings, pick-up trucks were strongly associated with republican voting districts, whereas sedans were more popular in democratic districts. The results are published in the journal Proceedings of the National Academy of Sciences.

Related: You Can Now Go Inside an Active Volcano With Google Street View

But the Street View cars revealed more than political leanings. Consistent with previous research, the study also showed associations between types of vehicles and race. For instance, Hondas and Toyotas most strongly indicated an Asian neighborhood. Cars manufactured by Chrysler, Buick, and Oldsmobile were associated with African American neighborhoods. And pickup trucks, Volkswagens, and Aston Martins were often associated with Caucasian residences. Income and education levels were also tied to neighborhood automobiles.

"This kind of social analysis using image data is a new tool to draw insights," Gebru, told The New York Times.

It's important to note that the data just shows an association, rather than a causation, study co-author Jonathan Krause, who works at Google Brain, told CityLab, The Atlantic's urban-focused news outlet.

"So it's not that you drive a pickup truck, therefore you are a Republican," Krause explained to CityLab.

In fact, making individual assumptions like that could have a great impact, he points out.

"There is a growing recognition in the field that your algorithm is only as unbiased as the data that you give it," Krause said. "The wrong way to use our study is applying it to an individual level, which would be dangerous to do."