NSA spying is causing Americans to self-censor their Internet activity

An MIT researcher claims he’s quantified some of the troubling self-censorship civil liberties advocates worried would result from public knowledge of mass spying.

The new study reports that Google users were slightly less likely (2.2 percent) to use search terms that the National Security Agency flagged as potential national security threats.

“This study is the first to provide substantial empirical documentation of a chilling effect, both domestically in the shorter term and internationally in the longer term, that appears to be related to increased awareness of government surveillance online,” explains the paper.

Compared to a semi-random list words, those that appear on the government watch list took a slight dip, noticeably around the time of Edward’s Snowden’s now famous leaks. This jives with other changes in behavior, such as increased use of anonymous search engine DuckDuckGo.

If readers buy the researchers argument, the effect is quite tiny. Screenshots from Google Trends reveal that, to the naked eye, it’s nearly impossible to see any difference in the volume of certain search terms before and after the leaks (the graph here is for the phrase “dirty bomb”).

It’s also important to remember that at least a few surveys find that a majority of Americans actually support the bulk collection of data.

But the research itself brings up an important philosophical debate: How much self-censorship should we tolerate before we care about the effect of mass surveillance? Snowden himself once responded to a question I posed about the “why should we care” question.

“Study after study has show that human behavior changes when we know we’re being watched. Under observation, we act less free, which means we effectively are less free,” he wrote to me.

In addition to the potential to fuel an authoritarian state, the most likely outcome was increased self-censorship, as citizens stifle their own speech in light of an ominous government surveillance apparatus.

But these programs have trade-offs. Increased spying could, indeed, save dozens (or more) lives by preventing terrorist attacks.

What level of self-censorship meaningfully impacts our democratic system?

It would be easier to say that self-censorship was a big problem if certain terms had nearly disappeared from Google (say because 60 or 70 percent of citizens stopped using them). But 2.2 percent?

If this isn’t compelling enough to show that citizens are censoring themselves, it is any longer a valid criticism against the NSA? And, if it is, why and what number would be sufficient? 1 percent? 0.0001 percent?

We live in a world of trade-offs. Beyond the knee-jerk defense or rejection of surveillance, it’s much better to get specific about what we lose on either side of the debate. This research is an important step forward past the usual talking points.