Using Trustworthiness to Weight Sources in Search Algorithms
Article: http://www.nationalreview.com/corner/414623/google-bias-search-engine-based-facts-wesley-j-smith
Throughout the past couple weeks in class, we have paid particular attention to search algorithms such as Google Search and their methods of displaying results for a given search query. The goal is simple: to order pages in terms of significance and relevancy. Working towards this goal, however, is not quite as straightforward. As this article addresses, incoming links from outside webpages are one of the most important factors when ranking pages. But not all incoming links are created equal. As we’ve discussed in class, when weighting the significance of incoming links, we must evaluate the credibility and importance of the sources of these incoming links.
This article explains just one way in how Google plans to evaluate the credibility of incoming links and the webpages they come from. In the near future, Google may parse websites and compute whether they contain any blatant factual errors. If a website contains any factual errors, its credibility would immediately drop and the weight of the websites it links too would as well.
While this method of weighting the trustworthiness of websites might help to better filter and rank search results, the article suggests that this might introduce a new source of bias to Google’s search results. For example, a relevant and trustworthy creationist website might state that the Earth is only 3000 years old. A search engine which considers the factual accuracy of a website might flag this as a factual inaccuracy and reduce the weight of a website even though the website accurately reflects the beliefs and views of creationism.