Machine learning systems can adopt its creators' prejudices. Researchers at Princeton University discovered how machines can also exhibit human bias because they reflect their creators.