Bias Warning from Artificial Intelligence Expert
When we look at the definition of artificial intelligence, we see that they are machines and software that can perform operations that the human mind can perform.
When we look at the definition of artificial intelligence, we see that they are machines and software that can perform operations that the human mind can perform.
Artificial of intelligence When we look at the definition, we see that they are machines and software that can perform operations that the human mind can perform. Amazon Alexa There are many different types of artificial intelligence, from those we see in our daily lives to Google's system that predicts our searches.
An artificial intelligence expert drew attention to a problem that needs urgent attention. Assistant professor of artificial intelligence at Leiden University Peter van der Putten, revealed a flaw in artificial intelligence.
The problem of bias in artificial intelligence

Decision making often occurs as a result of a complex process. achieved a single output Although it is (the decision taken), there may be different underlying factors. When evaluating these factors, observers according to their own experience they move. A recruiter's candidates "“Is he from the same school as me?” or "Who is not from my school?"" is a way of separating bias is the indicator.
Van der Putten At this point, he intervenes by saying that artificial intelligence is not good or bad on its own. Yet people your prejudices And their biases He also states that these systems can copy.
In addition, the artificial intelligence expert also works because of the data. your prejudices And your mistakes states that it can be transferred to systems. For example, a program developed for human resources artificial intelligenceIt was revealed that she copied the professional promotion problems experienced by women as a prejudice. Artificial intelligence trained based on other employee data, at higher levels When faced with equal candidates, he tended to reject female candidates.
Artificial intelligence learns bias

Although the bias inherent in rules and data cannot always be named directly, it can be revealed through statistics. This statistical Artificial intelligence, which discovers the structure in advance, then makes it a part of its own code.
co-founder of Apple Steve Wozniak There was another person who stated that algorithms may have developed prejudices against women. The company's credit card is available to men and women with similar credit characteristics. different limits had presented.
According to experts, artificial intelligence is as important as the data used to establish it. prejudiced and it is biased. It is not possible to prevent this, but it is possible to keep its effect controllable. Artificial intelligence It seems like as we learn something with new data, we learn something too.
Every business is unique. Let’s explore how we can tailor these ideas to your own digital project.