Monday, October 20, 2014

A Critique of Gillepsie's essay "The Relevance of Algorithms"


The Relevance of Algorithms is an essay written by Tareton Gillepsie, an associate professor of both Communication and Information Science at Cornell. He authored Wired Shut: Copyright and the Shape of Digital Culture, and has a lot of relevant information society experience in both academia and the industry itself. He is the co-founder of Culture Digitally, a blogging service. With assistance from his colleagues and the European Institutes for Advanced Study Fellowship Programme, he wrote this essay on algorithms in the information society for both an academic and professional audience. I believe that his piece was generally received as credible by its audience, because Gillepsie has since published many other peer-reviewed publications. However, UW library databases did not yield any scholarly reviewed results, nor did Google list any essay reviews in its search results (perhaps the algorithm intentionally filtered them out). A list of other pieces on this topic with varying conclusions is listed after my in-depth description of Gillepsie’s essay.

              Gillepsie’s  piece is comprised of sophisticated vocabulary one might expect an Ivy-League Professor to use and it explores the significance of algorithmic search engines, such as Google, in terms of our economic and political society. His main claim is that an algorithm is a knowledge logic, or a medium in which information can be explained from a certain, biased perspective. The example he used that I found most helpful is that an editorial is also a knowledge logic in so far as it conveys information from a biased perspective of an author. But how can an algorithm be biased? It is not a human with emotions and it simply produces certain outputs from certain inputs. First, the algorithmic formula may censor certain things out (something very prevalent in China which has many political repercussions). Second, if one were to search a qualitative query like “most athletic athletes,” it is inherently impossible to produce results without using judgments. Is strength the primary factor to be considered? What about hand-eye coordination, etc. For these two reasons, Gillepsie claims algorithms are knowledge logics.

                Subconsciously, internet users choose to use the search engine that is the least biased, according to Gillepsie. For example, if 9 out of 10 results of a particular search engine were advertisements, people would naturally switch services. For this reason, companies like Google attempt to gain credibility by saying that their formulas are the most objective. This is very hard to verify because the formulas are proprietary and kept secret. At the same time, Google sells advertising space. Gillepsie asserts that the relationship between users and service providers is entangled because changes in user groups affect algorithms (through advertisements) and changes in algorithms affect user’s server preference.

From the idea that an algorithmic search engine is biased, Gillepsie describes its implications to the online community, which he coins the networked publics. Search engines form a calculated public by using algorithms to determine what appears atop a query. In basic terms, algorithms and Google decide what is popular in networked publics. A good example of this is Twitter’s use of the hashtag to show what is trending in different areas. One thought I had is that the exposure caused by being trending may inspire more attention and might produce more popularity within the calculated public. Gillepsie says that there is friction between the networked publics and the calculated publics and one needs to account for this in assessing the societal impact of the internet.

It is evident that commercial institutions and government censorship impact algorithms, search results, and the network society. Gillepsie cautions against the temptation to make conclusions about the online society (networked public) based on the tremendous amount of data provided by search engines because it is inherently biased. At last, the political ramifications of algorithms online revolve around government censorship, while the economic implications are largely concerned with advertisements. These are the main points in The Relevance of Algorithms that I found most relevant to the information society.

 

Other similar pieces on the topic that may have reached different conclusions:

·        Media Rituals: A Critical Approach by Nick Couldry

·       Counterculture to Cyberculture: Stewart Brand, the Whole Earth Network, and the Rise of Digital Utopianism by Fred Turner

·        A Private Sphere: Democracy in a Digital Age by Zizi Papacharissi

·        The Business and Culture of Digital Games by Aphra Kerr

               

No comments: