# Why and How do algorithms exhibit biases?

The reason algorithms exhibit biases is that most of them have been trained on data that largely consists of biased information. Mrs. Rutherford and Mr. White argued that when searching for images on Google, most of the results appear to feature white hands. This isn’t, in my opinion, surprising in a country that is predominantly inhabited by white people.!


It was quite contradictory when I first heard that people are trying to call for unbiased algorithms, considering the historical context of this country, which is marked by racism, robbery, and chaos. These concepts are so deeply rooted in people’s minds that whenever they see something, such as a black person robbing a store or an Asian woman being sexualized on TV, they assume this is the norm of American culture. How can we have a free and fair algorithm under the pretext of stereotypes and stigmas?

In addition, unlike humans who can discover mistakes and logical flaws in our own arguments, cold-minded machines are very different in terms of error detection. When it comes to politics, cultures, and societies, algorithms cannot operate and thoroughly evaluate situations like humans do. While ChatGPT can solve a mathematical problem within less than a second, there are many gray areas that machines cannot account for when it comes to socio-political subjects. Therefore, expecting AI to think like humans is still a naive expectation.

In the near future, however, I cannot rule out the possibility that AI will be able to generate results that are more impartial and represent the comprehensive side of society. But until this point, we cannot ask AI to do more than just answer some complicated questions.