On Monday as I was reading reactions on the web about the successful raid on Osama bin laden’s compound, I kept coming across a quote “I mourn the loss of thousands of precious lives, but I will not rejoice in the death of one, not even an enemy. – Martin Luther King, Jr.”   Part of me felt it was a beautiful sentiment.  Part of me thought it was an impossible standard.  And I am always a bit worried about the borderline for these kind of statements between sincerity and Hallmark. In retrospect, I am happy that I didn’t hit the “reblog” button on tumblr.  Why?  It turns out the quote is not really a Martin Luther King quote.  Instead, it was an introductory sentence to an actual MLK quote.

The spread of the fake MLK quote is a great example of an information cascade.  This topic has been on my list of possible blog posts for a while and Monday’s rapid fire spread provided a terrific reminder of how important this phenomenon has become.  An information cascade occurs when people are spreading (mis)information based on observing the actions of others (instead of verifying the underlying information).  Most people hit “like” or “reblog” or “retweet” on this quote because others had done so before them, instead of looking for the source.

Pre-Internet, a classic example used to illustrate information cascades was competition between neighboring restaurants.  In the absence of additional information, there will be a huge difference between two restaurants that are next to each other, solely based on who attracts the first customers of the evening.  As further potential customers arrive, they will be more likely to go to the restaurant that already has guests.   In the extreme, at the end of the evening, the restaurant which had the first guests will be full and the other empty.  If you are not convinced of that as a theory, just walk by a set of restaurants in a tourist part of town early for a mealtime to see how aggressively owners are trying to attract the first few guests!

Now the fascinating thing about the Internet is that it has in theory made it easy to take a verification step and check source information instead of cascading based on others’ actions.  So in the restaurant example you can just pull up Foursquare or Yelp to get information.  But the Internet has also made it that much easier to spread unverified information with a single click (reblog, retweet, like).  The state of the art today is that spreading is easier than verifying which means that we are getting more, not fewer information cascades.  That is especially true because with social networks we are observing the actions of friends or at least people we know (instead of random strangers) and are thus more likely to copy their actions.

What can be done about it?  First, we should all remind ourselves daily of the danger of information cascades.  For instance, I tend not to retweet anything with a link in it without first clicking through.  Second, we should try to create technology support for verification.  A universal “check this” button might be very useful thing.  I don’t expect it to necessarily be powered by something fully automated – it might be enough to create a list of things that people have asked to see verified to peer produce some checks (an opportunity for Wikipedia?).

On Monday as I was reading reactions on the web about the successful raid on Osama bin laden’s compound, I kept coming across a quote “I mourn the loss of thousands of precious lives, but I will not rejoice in the death of one, not even an enemy. – Martin Luther King, Jr.”   Part of me felt it was a beautiful sentiment.  Part of me thought it was an impossible standard.  And I am always a bit worried about the borderline for these kind of statements between sincerity and Hallmark. In retrospect, I am happy that I didn’t hit the “reblog” button on tumblr.  Why?  It turns out the quote is not really a Martin Luther King quote.  Instead, it was an introductory sentence to an actual MLK quote.

The spread of the fake MLK quote is a great example of an information cascade.  This topic has been on my list of possible blog posts for a while and Monday’s rapid fire spread provided a terrific reminder of how important this phenomenon has become.  An information cascade occurs when people are spreading (mis)information based on observing the actions of others (instead of verifying the underlying information).  Most people hit “like” or “reblog” or “retweet” on this quote because others had done so before them, instead of looking for the source.

Pre-Internet, a classic example used to illustrate information cascades was competition between neighboring restaurants.  In the absence of additional information, there will be a huge difference between two restaurants that are next to each other, solely based on who attracts the first customers of the evening.  As further potential customers arrive, they will be more likely to go to the restaurant that already has guests.   In the extreme, at the end of the evening, the restaurant which had the first guests will be full and the other empty.  If you are not convinced of that as a theory, just walk by a set of restaurants in a tourist part of town early for a mealtime to see how aggressively owners are trying to attract the first few guests!

Now the fascinating thing about the Internet is that it has in theory made it easy to take a verification step and check source information instead of cascading based on others’ actions.  So in the restaurant example you can just pull up Foursquare or Yelp to get information.  But the Internet has also made it that much easier to spread unverified information with a single click (reblog, retweet, like).  The state of the art today is that spreading is easier than verifying which means that we are getting more, not fewer information cascades.  That is especially true because with social networks we are observing the actions of friends or at least people we know (instead of random strangers) and are thus more likely to copy their actions.

What can be done about it?  First, we should all remind ourselves daily of the danger of information cascades.  For instance, I tend not to retweet anything with a link in it without first clicking through.  Second, we should try to create technology support for verification.  A universal “check this” button might be very useful thing.  I don’t expect it to necessarily be powered by something fully automated – it might be enough to create a list of things that people have asked to see verified to peer produce some checks (an opportunity for Wikipedia?).