Personalization versus ‘Filter Bubble’: The influence of personalization on the quality of search queries
Before the mid-2000s search engines find search results using conventional keyword search and ranking techniques. In that period it does not matter who is searching or what location search happens everybody gets the same results if used the same set of keywords. It does not matter that search results are relevant to the user. Personalized search is a significant upgrade to conventional web search. Personalized search uses more factors to search not just keywords and gives more personalised and relevant results to users. As an example, someone search pizza at two geographical locations result is different for each other.
The personalised search came to the picture in 2004. In 2004 google introduced personalised search and implemented to Google search in 2005. Until 2009 its only applied to users who logged into Google accounts. Afterwards, it applies to all users who use google. Although technical details of these commercial systems are proprietary, however, there are few techniques google use to personalise the search. Search results can be different based on location. For example, search results for word “restaurants” can be different in different geo-locations. Users location can be obtained using location service or IP address. Also, search results get personalised due to device user tries to search. It also can be affected by previous searches and frequent site visits. Also search engines in social network platforms like Facebook or LinkedIn, personalisation could be achieved by exploiting homophily between searchers and results. As a conclusion, these search engines take advantage of peoples information and give more personalised and relevant search results for given search query.
The filter bubble is a result of the personalised search. Eli Pariser introduces the concept of the filter bubble in his book “The Filter Bubble” in 2011. The filter bubble is own personal unique universe of information that person live online. The filter bubble is a result state of websites algorithm selectively decides according to some parameters like location, and previous search results and users become separated from viewpoints that they disagree and isolate on their ideological world. What is on ones’ filter bubble depends on who they are and what they do. The person does not get to decide whats get into their filter bubble or what should filter out. This drive an individual to a state of intellectual isolation. Facebooks personalised news feed is one good example of the filter bubble. In the beginning, the internet was about to exchange diverse ideas freely among everyone. The filter bubble threatens to the very foundation of the internet.
There are many benefits and uses associated with personalised search. One of the most crucial benefits of personalised search is that it gives more relevant search results than conventional keyword web search and ranking over to quantify the importance of it. Personalised search helps save time by giving the most relevant results to its user. Find the most relevant information from a search query is tough work because the search engine has to go trough billion of web links and terabytes of data to get it. Therefore increment of the relevance of search results and reduce the ambiguity of search results are one good advantage get over the personalised search. Another good result of the personalised search is, it gives locally more relevant search results. It is a good situation for local businesses. When someone searches “restaurants” it gives geographically near restaurants to the searcher. It is more relevant to the person who searches as well as the local businessman. Personalised search indirectly markets local businesses. In a way, it is good to stick in a filter bubble because it is a less distractive environment. Students can keep the focus on studies without unwanted distraction. In a way, it gives a perfect version of a customised internet for us. Therefore it can conclude the personalised search gives more relevant search results to search queries, but it stuck us in their perfect world for us.
There are two sides on every coin, the same way personalised search has downsides as well. Personalised search makes our lives more comfortable, but in order to do that, it requires our personal information such as previous search history and location. There is a tradeoff between our privacy and relevant quality search results that need to consider. It is one downside of personalised search. Not only that but also these giant internet companies use our private information on target advertising. Also, the filter bubble violates the most fundamental ideology of the internet. It is meant to share various ideas with everyone. However, these programmed gatekeepers violate that basis and isolate us on information bubble they thought that fit to us. Users are the one that needs to decide what are the things that need to see, not some algorithms. Filter bubble disrupts the democracy too. In this century the internet plays a significant role as a communication media. In the democracy, it requires everyone to be informed equally. In election time, it needs to consider every parties opinion before making a decision. However, users stuck in an isolated environment that computer algorithms decide what should we see know and what shouldn’t. It is a critical problem, and whoever sits behind the giant internet companies can control and manipulate the public opinion. Therefore it can conclude that the personalised search has downsides as well. Filter bubble threat to the very foundation of the internet, democracy and privacy too.
As a conclusion, personalised search makes our life a lot easier and comfortable by giving more relevant and less ambiguous results. However, personalised search let to a new problem known as filter bubble. State of computer algorithms decides what should users see and shouldn’t and let users to intellectual isolation. Filter bubble has terrible effects on privacy, a challenge to democracy and its threat to the very foundation of the internet, Share a broad range of ideas among everyone too. So it is good to use the personalised search, but users should be the ones who decide what should they see or not.
References
* https://en.wikipedia.org/wiki/Personalized_search
* https://en.wikipedia.org/wiki/Google_Personalized_Search
* http://tentacleinbound.com/articles/personalized-search
* https://en.wikipedia.org/wiki/Filter_bubble
* https://arxiv.org/pdf/1605.04624.pdf
* https://searchengineland.com/study-asks-can-you-trust-googles-personalized-search-results-64709
* https://www.technologyreview.com/s/423596/how-useful-is-personalized-search/
* https://fs.blog/2017/07/filter-bubbles/
* https://ijournals.in/wp-content/uploads/2017/07/IJSHRE-2441.pdf
