AI bias

AI bias appears when a system produces systematically unfair, skewed or discriminatory results toward certain groups of people. The cause usually lies in the training data ✕ Training data is the set of examples an AI model learns from. Its amount and quality determine how good —and how biased— the model will be. More in the glossary → : if it reflects historical prejudice, is incomplete or overrepresents one group, the model learns and reproduces those distortions. Documented cases include hiring, credit approval or facial recognition that worked worse for particular groups. The problem is delicate because AI ✕ Artificial intelligence is a computer system’s ability to perform tasks we associate with human intelligence, such as understanding language, recognizing images or making decisions. More in the glossary → carries an appearance of objectivity ("the algorithm says so") that can hide the bias. Reducing it requires more representative data, specific fairness testing and human oversight. It is one of the central ethical issues when using AI for decisions that affect people.