Technology Experts Warn of Hidden Racial Bias in AI Systems

In 1998, the author unintentionally created a racially biased AI algorithm.
Recent failures of AI chatbots and image generators have revealed inherent racial prejudices.
Theories of white privilege help explain how well-meaning scientists create biased AI without realizing it.
Bias persists undetected deep within AI systems, even in modern machine learning.
Diverse teams are needed to design AI and avoid prejudice.