I just finished reading Better: A Surgeon's Notes on Performance by Atul Gawande. It's my book club book for this month, and although I didn't think I would find it that interesting, I was surprised by what I learned.
In the first chapter, Gawande describes how health workers in Karnataka, India responded to a polio outbreak in June 2003. Their task was to inoculate 4.2 million children in a matter of days to contain the one "index case" that had been identified. Eradicating polio is all about the last drop of effort--the immense expense and difficulty of containing outbreaks as soon as they occur. He describes it as a challenge of diligence--of the kind of commitment to do what needs to be done in the face of depressing odds and often impossible conditions. But the doctors do it because they know that this is the only way to win and hopefully add polio to the list of plagues that has been wiped from the face of the Earth. In addition the the Herculean effort involved in vaccinating millions of children around the epicenter of an outbreak, he also notes the information challenge--the problem that every hour that goes by before the outbreak is noticed, the scope of the mop up operation increases exponentially. If we could instantly recognize that an infectious disease was breaking out, the mop up would be orders of magnitude easier. And when we fail to detect the outbreak...it jeopardizes decades of work to eradicate the disease.
The solution to the information problem may be coming...from Google, of all places. Several years ago, the Public Health Agency of Canada, using technology developed by Nstein, Inc., developed something called the Global Public Health Intelligence Network (GPHIN). The system monitors tens of thousands of news feeds from all over the world, performs sophisticated text mining to extract contextual meaning, then watches for patterns that could indicate an infectious disease outbreak. An example could be a pattern of news reports in a region about people with disease symptoms. The system would give medical professionals immediate notice of a potential crisis developing. The GPHIN system was instrumental in containing SARS in 2002.
This technology caught the attention of Larry Brilliant, who described it as the nucleus of what he would like to develop if he had the resources... Brilliant won a 2006 TED prize to help implement his vision, and assumed leadership of Google's philanthropic division, Google.org. Since 2006, there has been little public information about the progress of this project, but one can only imagine what will emerge from Google Labs to take an awesome idea to the next level.
When we begin to think about the possibilities behind massive parsing and text analytics applied to public media, it is easy to think of a dark side...or at least a more business-oriented, less philanthropic side. I wonder if my blog is being parsed by DoD computers in search of patterns that could indicate I am secretly coordinating an Al Qaeda cell? Are my phone calls being converted to text and similarly parsed? I'm not that worried; to me it seems the positive value is probably so much greater than any conspiracy fantasy I might entertain.
In Malcom Gladwell's Blink, he described how, in World War II, the British assembled thousands of "interceptors" to listen to encoded morse code transmissions from German operators. Although they could not decode the transmissions, they could, by listening to the cadence of the transmissions, identify the individual "fists" of the Germans. They know who was sending the messages. Over time they were able to use this information to follow the movements of different units.
Text analytics is very different than simply identifying who is talking. But it is a similar step toward automation of a subconscious skill...not so much the development of artificial intelligence, but of a catalyst for collective intelligence. James Surowiecki talks of the great challenge in harnessing the power of the wisdom of crowds as building structures and systems to effectively aggregate the crowd's intelligence. The kind of text analytics, early warning system envisioned by Larry Brilliant is an evolutionary leap forward--not just for the case of identifying pandemics, but for accelerating the feedback loop in all sorts of social improvement.
Most of what I've written here today is simply stitching together a few fascinating articles and other people's blog posts. But imagine if a Google technology were crawling the entire blogosphere looking for patterns and connections...and imagine you could ask such a system a question. I feel we are venturing into serious "psychohistory" territory here. Rather than fearing the future, we can imagine a world where our many contributions, however ever individually small, might be pulled into a greater good.
http://blog.davewrites.com/index.php?title=saving_...