Blog

  • Review of Eli Pariser's "The Filter Bubble"

  • The ‘Filter Bubble’ by Eli Pariser carves out a new space in the growing body of literature that is being coined ‘internet activism’. Eli Pariser’s new book describes the phenomenon in which websites use algorithms to selectively guess what the user would like to see based on information about the user. Pariser’s prime examples of the ‘filter bubble’ are the personalised search results of Google, and the personalized news stream on Facebook. He claims that the user gets less exposure to information or viewpoints that are more diverse than their own. 

    The formal definition of the ‘filter bubble’ provided by Pariser is a ‘personal ecosystem of information that’s been catered by these algorithms’. He warns that the filter bubble will have detrimental effects as it ‘closes us off to new ideas, subjects, and important information...creates the impression that our narrow self-interest is all that exists’. 

    What comes out of the critical discourse surrounding this this book is that Pariser has over-estimated the extent to which any user currently experiences a filter bubble (see Jacob Wesiberg’s non-scientific experiment in Slate Magazine).  However, it would be naïve to think that Google (and others) do not hold the intention of increasing the amount of personalization in their services. This would have been fortified in the wake of its newest incarnation, the infant terrible of the social networking world, Google+. 

    You can sympathise with Pariser’s fear that the internet will no longer be an open tool to serendipitously encounter new knowledge; to have a popular voice making a noise about this will do something to inhibit that fear being realised.  

    However, his treatment of the internet as a single phenomenon tends to conflate different aspects of the web. Should ads be jumbled in with a social media newsfeed? You find that users respond positively when they meet advertising that is relevant to them. The word ‘relevant’ is defined as “having a bearing on or connection with the matter at hand”. The detrimental effects of the ‘filter bubble’ are defined as being able to ‘‘close us off to new ideas, subjects and important information’. 

    When these two definitions are put next to each other you can start to see how the ‘filter bubble’ breaks down when applied to advertising. Fundamentally, people are not online to consume advertising; it is a by-product of getting all these websites for free.  Most people are happy then to receive advertising that is relevant rather than irrelevant when they are online. Wouldn't it be great if every single ad had a 'baring or connection' with your interests!

The concept itself, of a digital ‘filter bubble’, is no doubt one that will eventually enter the general parlance of the tech community. However Pariser limits himself, by having a general problem with algorithms curating the web. He doesn’t allow himself to see that in some places, creating a more relevant user experience is a positive step for the development of the internet, and here the role of algorithms is essential.