It is not the first time that Google’s search process algorithms have been controversial. In 2019, the company had to correct the functioning of the intelligence that associated the word “Lesbian” with pornographic content. French news site Numerana has noticed the pattern. In a conversation with the Gizmodo website, a Google spokesman said at the time that the company “developed an algorithmic solution to provide high-quality results not just for that query, but others of the same kind.”
Something similar happened in Brazil last year. The expression “macho escroto (scumbag)”, when typed in the Google search bar, pointed to a photo of ex-BBBs Hadson, Lucas, Prior, Guilherme and Petrix. The participants caused many intrigues inside the house in the last edition of the program.
Now, several users of the site point to a new incongruity. This is because, when searching for the “mentally ill flag”, LGBT and trans flags appear in large numbers. So far, Google has not officially commented on the repair of the algorithms for this case.
Understand how Google’s search algorithms work
With the amount of information available on the web, finding what you need would be next to impossible without any help in classifying it. Google’s ranking systems organize hundreds of billions of web pages in the search index to provide the most useful and relevant results in a split second and present them in a way that helps you find what you’re looking for.
These classification systems are composed not by one, but by a series of algorithms. To provide the most useful information, the search algorithms analyze several factors, including query words, relevance and usability of the pages, knowledge of the sources, as well as their location and settings. The weight applied to each factor varies according to the query nature.
For example, updating content plays a more important role in responding to queries about current news topics than about dictionary definitions. To help ensure that research algorithms meet high standards of relevance and quality, there is a process that involves real-time testing and thousands of trained research quality external evaluators worldwide.
These quality assessors follow guidelines that define our goals for research algorithms and are publicly available to anyone.