National Public Radio recently featured a story about a new possibility for United States intelligence gathering. It seems the CIA has finally uncovered the same technique (by now, almost age-old) used by everyone else in the world in need of answers – just Google it.
It’s a comical oversimplification, of course. Google Analytics is being looked to as a new form of open-source intelligence to analyze the atmosphere of groups of people. The idea is that nations already use the Internet as a tool to cultivate social interests, which is activity that can be tracked and measured in a systematic way.
Indeed, we Google everything, from our most casual interests to our most frantic problems, the movies we watch, the symptoms of diseases we might have contracted, movements of social or political interest and everything in between. The Internet is the evolving, real-time imprint of society, and the ability to identify what forces are animating Internet searches would allow governments to better predict and respond to things like protests or disease outbreaks.
If you think about it, the theory makes sense. We pour our souls out into the Internet and social media, and rely very heavily on search engines like Google to tell us we need to know about very nearly everything. Our Internet searches tend to be much more frank than our face-to-face conversations – to get good answers, you have to be completely honest – and in that way, you could claim that Google knows you better than your closest confidants ever will.
In 2009, Google started Google Flu Trends, which tracked when people Googled their flu-like symptoms, and often proved days and weeks more accurate than government predictions in forecasting outbreaks.
In the case of the recent unrest in Tunisia, researchers looked at Google Trends for terms like “Tunis” to see how much attention Egyptians were paying to the events there. The question posed now is, could Google Trends have actually predicted the subsequent protests in Egypt? Is it possible to read an entire population based on how they traffic the Internet?
Wiretaps and espionage are still effective for listening in on conversations between individuals, but society at large tends to act in a more candid fashion. The Internet, which is the umbrella for news, blogs, social media and all the separate forms of information trading we like to specialize in, is effectively sounding board for all the world’s feelings and opinions. Before big things happen, people have to talk them out. The Internet acts as the social boiler, and the idea is to be able to measure when the pressure begins to build too high.
Google Analytics has been active for several years now, tracking and cataloging search terms as Google Trends. Market researchers have been using targeting relevant products to people who are probably interested in them. I was once shopping online for a new camera, and after I had done a few searches for the model I began noticing that the very camera I was looking for started showing up in ads around my regular Internet hotspots. Conveniently enough, I ended up getting it for a really good price. Rest assured, this was no coincidence.
But how far can these clever tactics be taken? If Google can suggest to me what camera I should buy, can it make other, more intimate predictions? By cross-referencing the search terms we use and the kinds of web pages we visit, it’s not unreasonable to think that a computer could make probable conclusions about the mood or intentions of a user.
The thought reminds me of the old Microsoft Office Paperclip assistant, who would constantly judge your work and say irritating things like, “It looks like you’re trying to write a letter.”
“It looks like you’re trying to instigate a coup d’état.”
Let’s take this to a deeper level. Could a computer use my search terms to make calculated assumptions about who I am as a person? We pour out our souls into the blogs we read, the kinds of videos we watch, the news we respond to, and generally the ways we spend our time online. Could a computer read what I type, see what I see, and use simple algorithms to determine the probability that I’m, say, planning a party, or interested in the Civil War, or especially passionate about a present issue in politics?
Could a computer even look at my browsing habits and determine that I might be clinically depressed, or maybe a suicide risk?
It’s all very Big Brother, to be sure, but the possibilities are fascinating nonetheless. Our generation has watched the Internet evolve boundlessly, almost as living organism, becoming faster, smarter and more complex all the time. What will it be capable of in the next generation?
