The Dangers of Invisible Filtering Algorithms

Do you ever wonder how Facebook decides which items are included in your news feed and which are excluded?  Presumably, complex algorithms use information about your past use of Facebook (things you click on, things you "like", etc) to determine what items are most important or relevant to you.  The most concerning element of this is that you don't know, and you have no control over what is being excluded from your feed.

In 2002, M. T. Anderson wrote a novel called Feed that predicted a future in which everyone is connected to the "feed", which constantly makes suggestions for goods and services that someone may be interested in based on their consumer profile.  The idea behind it being that corporations would have an easier time increasing consumption and predicting demand by simplifying  people into basic consumer archetypes.  Ten years later, it appears that Anderson's predictions may have been surprisingly accurate.

Most of us are at least aware that Facebook is filtering our news feeds, But what is much more alarming is that Google search results are also heavily influenced by your past browsing habits.  In fact most of the busiest internet sites use this type of filtering to some degree.  Eli Pariser has recently written very informative book called The Filter Bubble which discusses the issue of these filtering algorithms in great detail.  In a recent TED talk he talks about the book, and how two different users performing a Google search on something as basic as "Egypt" can get vastly different results.

One of the dangers of this type of filtering is that people may have a tendency to click on more entertaining and less informative links most of the time, and this may lead to the more informative type links being left out of your searches or feeds entirely.  The simplifying effect it may have on our information sources could be very significant, as it tends to supply give the type of information that you seem to prefer, thus isolating you from alternative ideas, viewpoints or people you seldom interact with.

Perhaps due to the recent negative attention, just last month Google launched Search, plus your world which gives users more control over what factors effect their search results.  Most importantly, it includes an "opt-out" feature which allows you to search without any on-line social networks or user history effecting the results.  We need to encourage this type of transparency and increased user control in user interfaces if we want to keep the internet a source of vast and free information, instead of a means of self-indulgence that just further simplifies and tunnel blinds its users.

In Naomi Klein's No Logo  she talks about how popular clothing stores like the Gap and Old Navy started out trying to understand youth culture and determine what products would be considered in-style.  Before long this relationship switched and teenagers were turning to these stores to find out what was in style, and the stores themselves were having a significant influence on youth culture.  I believe a rather disturbing analogy can be drawn between what these chain clothing stores did in the 90's and what is currently happening on some of the most heavily used internet sites.  At what point does it change over from these algorithms trying to determine what information we want, to these algorithms determining what we're interested and informed about.