Saturday, March 26, 2011


Imagine your grandfather, or even your father, trying to decipher your Twitter tweets or Facebook status updates. See how things could get problematic quickly? Don't worry- your dad isn't the only person with the problem- it appears American intelligence agencies had some trouble, too, when it came to predicting the revolutions in the Middle East. But next time, instead of asking their teenage daughters for help, open source intelligence analysts may be focusing on trends in Googles and Tweets instead of the unique content.

While many have dubbed the events in Egypt the "Facebook Revolution" due to the amounts of video posting, information sharing, and relative-contacting done online, early stage event planning was also evident on the social networking sites. While inviting your friends to an event will certainly help disseminate the information locally, in countries where Internet penetration is extraordinarily high - like Korea's 81%- sifting through thousands of statuses about cute puppies and Charlie Sheen for a tweet on the next big thing is nearly impossible. Thus feed aggregators and trend monitors like Google and Twitter Trends may be the best options for monitoring situations in the pre-event and initial-event.

For example, have you ever searched online for flu symptoms or even just the word "headache"? If so, you may have helped Google in their trend monitoring software- Google Flu Trends- which takes keywords and phrases they believe to be good indicators of flu activity, and then plots the results by state and country. Google often beat the NIH on predicting flu outbreaks by a week or more.

Gabriel Koehler-Derrick, an instructor at West Point, used Google Trends to monitor the moods of the Egyptian people from January 25th onward, to see if Egyptians followed the revolution in Tunisia, and then compared the number of searches to things like "weather" and "Egyptian pop stars". The results suggested there was a definite increase in searches in the weeks leading up to their own revolution. Based on these results, Koehler-Derrick suggests we may be able to use current trends to predict the future.

From his interview with NPR News regarding the technology: "In other words, few [Egyptians] seemed interested enough in the Muslim Brotherhood to search for them on Google. So how much of a role could the group have been playing in day-to-day conversations in Egypt? ". (Note: Not surprisingly, thier influence at present, while still significant, is not currently anywhere near US intelligence community estimates.)

In actuality, intelligence agencies do hope to expand the Google Trends software for use in a Minority-Reportesque program known as Recorded Future. While the program is still very much in the initial phase, the software would search for the invisible links intelligence analysts might not. There is little reported on exactly how much/how many predictions are made, and in a community already flooded with information overload, it doesn't look like analyst positions will be going away any time soon. So, if you're applying for a job in the intelligence community after graduation, don't forget to add "Expert Googler" or "Intermediate Tweeter"- or perhaps even "Trend Follower" to your job skills.

No comments: