Once a pervasive surveillance infrastructure is in place, a government will always have an incentive to abuse that power, according to new research from the UC Berkeley School of Information.
Unlike most previous studies, the researchers used a game theory approach to understanding the competing incentives that drive both governments and their citizens.
“What happens if you really have a pervasive monitoring system?” asked visiting professor Paul Laskowski. “A lot of behavioral research has looked at individual decisions, but so far very little has focused on the overall cost to the system. We wanted to try to set off on a path to quantify some of these costs. Our paper is the very first attempt to have a look at that in an economic framework.”
The researchers developed a model that predicts how much a government will abuse its power and how citizens will react to that abuse, based on the level of surveillance infrastructure available and the overall effectiveness and popularity of the government. The study, by Laskowski, Benjamin Johnson, Thomas Maillart, and professor John Chuang, was presented at the 13th Workshop on Economics and Information Security, held last month at Penn State University.
With a popular government, some degree of surveillance may be beneficial, the researchers found, since it makes the government more stable and protects it against being overthrown. But what’s best for the government is not what's best for the country — the government will always want to conduct more surveillance, above and beyond what’s best for the country.
“If you implement a strong surveillance capability, it tilts the playing field,” explained Laskowski. “It tilts the incentives, so that governments have more incentive to be abusive. We think that relationship has been unexplored.”
If the government is unpopular, on the other hand, the availability of surveillance technology always makes the citizens worse off.
The study is “pure theory” according to the researchers. Rather than measuring actual levels of surveillance or corruption in a particular country, they instead constructed a general model to explain the interaction of incentives. That model could be applied to democratic governments, where the party in power risks being voted out of office by disaffected voters, or to governments that fear a popular revolt or coup.
Abuse or corruption could also look different in different contexts, the researchers said. Abuse could include spying on political opponents or rivals, suppressing dissidents, drone attacks, wiretapping, or any other way that a government uses surveillance systems to protect its own power.
“We have a really broad result that says that under general scenarios, surveillance technology creates the incentive for more corruption,” said Laskowski.
But it’s difficult to predict citizens’s reactions to a full surveillance state, because “the better the surveillance is, the more secret it is,” said Johnson. “When you have full surveillance, all the abuses are done in secret, and citizens aren’t unhappy — because they don’t know.”
“The government always has an incentive for more surveillance technology — but that’s not what's good for citizens,” said Laskowski. “For citizens, there’s some peak; once the government is safe and secure and we have the system that’s good for us, we don’t want more surveillance. At that point, the government becomes more abusive — and that’s to our detriment.”