The Google Algorithm Doesn’t Need to Change: We Do
When fake news is purveying Google’s feature snippet, only we are to blame.
Here is a list of presidents that were in the KKK: William McKinley, Woodrow Wilson, Warren G. Harding, Calvin Coolidge, and Harry Truman. At least, that is what our oracle Google said in its feature snippet when I experimentally searched “which presidents were members of the KKK?” None of these presidents were ever seriously considered to be members of the KKK by historians. However, in a ruthless post-truth era, alternative facts continue to spread and has now entered the annals of our beloved search engine’s feature snippet. The question now is, why?
Throughout the last few years the Google feature snippet has featured multiple (and mostly political) gaffes. These include declaring former president Barack Obama king of the United States, declaring Republicans to be modern day Nazis, and explaining why women are naturally evil. These issues have led to outcry among people claiming that fake news organizations are gaming Google’s algorithm on long-term search results.
Can Fake News Organizations Purposely Game Google’s Algorithm?
This is a nefarious claim, and if it were the case the algorithm would need a complete overhaul. However, I doubt it. Google’s algorithm is a complex system that is impossible to game without an immense amount of resources. This is due to Google’s dedication to constantly updating their algorithm and understanding user behavior in the event of cheating. Below is a filtered list of some updates that Google has made to the algorithm just within four years. (For the full list click this link here)
What these updates mean is that gaming the system requires going beyond just typing a few keywords repeatedly. Gaming the Google algorithm would require the collusion of a vast network of hundreds of thousands of people citing the article on blog posts and news networks. With that being said, I have a very difficult time believing that fake news companies call up individuals and other news companies in order to link articles.
Is it possible that fake news sites directly draw upon vast networks? Unlikely, without direct government interference. While social engineering (paying people to share or promote a link) is prevalent within social media, Google draws no influence from social media to determine its search rankings. Therefore, if there was a vast conspiracy, fake news sites would be forced to solicit links directly to articles from other sites and blogs. Soliciting links is a noticeable strategy and-unlike Facebook and Twitter-is in direct violation Google’s policy. This breach of policy would certainly be obvious to Google and result in these sites getting de-indexed with little chance of being re-indexed. This occurs even when governments directly interfere and even though they may reach the top of search results for a couple of days, it is unlikely they would be at the top long enough to warrant a spot in the feature snippet.
So how is this fake information floating around Google’s feature snippet for long periods of time? The answer to this is mainly partisan search biases which directly causes the spread of information through citations.
Most people, outside certain left fringes, would not normally google whether the Republicans are Nazis. At the same time, most people, outside certain right fringes, would not normally google Barack Obama being king of the United States. Google’s algorithm is currently built to provide the most relevant information, not necessarily the most accurate. Unfortunately, for those that search for Barack Obama’s supposed kingship, the information is biased towards the right spectrum because that is most relevant to the people who do these kinds of searches. The same applies to Republican Nazism across the left spectrum. These biased searches leads people within both sides of the political spectrum to cite the inaccurate information and therefore make it more relevant for those that search these individual questions.
Google’s algorithm is currently built to provide the most relevant information, not necessarily the most accurate.
Google has made several attempts to crack down on the creation and spread of fake information. One way has been to remove their display ads from fake news sites. This cuts revenue from fake news sites and thus reduces their incentives to create these articles. Google has also taken steps to hire quality raters to comb through search results to make offensive and inaccurate material less relevant. These are fantastic steps forward and will help add some degree of control to inaccurate and offensive information.
However, no matter what steps Google takes to move forward, some feature snippets will always slip through the cracks. Some bloggers are asking Google to remove their feature snippets. That scenario is unlikely since the feature snippets help give Google Home a competitive edge over Amazon Echo. On top of that, Google’s feature snippet has been a modern convenience and has mostly remained extraordinarily accurate.
The key to changing the Google feature snippets is for us to change ourselves and how we read information. We cannot allow ourselves to constantly have finger-on-the-trigger reactions to the absurdist posts that reach the spheres of the internet. It is our duty as humans to carefully question information and do further research to separate fact from fiction. Otherwise, what is the point of having information at the tip of our fingers if the information is constantly false?
Google has reached out their hand in order to combat fake news. Now it is our turn to reach out ours.