Algorithmic Policing

From P2P Foundation
Jump to navigation Jump to search



Christina Larson:

"Writing in the Washington Post earlier this year, Xiao Qiang, a professor of communications at the University of California, Berkeley, dubbed China’s data-enhanced governance “a digital totalitarian state.” The dystopian aspects are most obviously on display in western China.

Xinjiang (“New Territory”) is the traditional home of a Chinese Muslim minority known as Uighurs. As large numbers of Han Chinese migrants have settled in—some say “colonized”—the region, the work and religious opportunities afforded to the local Uighur population have diminished. One result has been an uptick in violence in which both Han and Uighur have been targeted, including a 2009 riot in the capital city of Urumqi, when a reported 200 people died. The government’s response to rising tensions has not been to hold public forums to solicit views or policy advice. Instead, the state is using data collection and algorithms to determine who is “likely” to commit future acts of violence or defiance.

The Xinjiang government employed a private company to design the predictive algorithms that assess various data streams. There’s no public record or accountability for how these calculations are built or weighted. “The people living under this system generally don’t even know what the rules are,” says Rian Thum, an anthropologist at Loyola University who studies Xinjiang and who has seen government procurement notices that were issued in building the system.

In the western city of Kashgar, many of the family homes and shops on main streets are now boarded up, and the public squares are empty. When I visited in 2013, it was clear that Kashgar was already a segregated city—the Han and Uighur populations lived and worked in distinct sections of town. But in the evenings, it was also a lively and often noisy place, where the sounds of the call to prayer intermingled with dance music from local clubs and the conversations of old men sitting out late in plastic chairs on patios. Today the city is eerily quiet; neighborhood public life has virtually vanished. Emily Feng, a journalist for the Financial Times, visited Kashgar in June and posted photos on Twitter of the newly vacant streets.

The reason is that by some estimates more than one in 10 Uighur and Kazakh adults in Xinjiang have been sent to barbed-wire-ringed “reeducation camps”—and those who remain at large are fearful.

In the last two years thousands of checkpoints have been set up at which passersby must present both their face and their national ID card to proceed on a highway, enter a mosque, or visit a shopping mall. Uighurs are required to install government-­designed tracking apps on their smartphones, which monitor their online contacts and the web pages they’ve visited. Police officers visit local homes regularly to collect further data on things like how many people live in the household, what their relationships with their neighbors are like, how many times people pray daily, whether they have traveled abroad, and what books they have.

All these data streams are fed into Xinjiang’s public security system, along with other records capturing information on everything from banking history to family planning. “The computer program aggregates all the data from these different sources and flags those who might become ‘a threat’ to authorities,” says Wang. Though the precise algorithm is unknown, it’s believed that it may highlight behaviors such as visiting a particular mosque, owning a lot of books, buying a large quantity of gasoline, or receiving phone calls or email from contacts abroad. People it flags are visited by police, who may take them into custody and put them in prison or in reeducation camps without any formal charges." ( )