Overview

AI bias is not usually intentional; it is often a reflection of existing biases in the real world that are captured in the training data.

Sources of Bias

  • Data Bias: Training data that overrepresents or underrepresents certain groups.
  • Algorithmic Bias: The design of the algorithm itself favoring certain patterns.
  • Human Bias: Biases of the developers who choose the data and define the goals.

Consequences

Can lead to unfair treatment in hiring, lending, law enforcement, and healthcare. Addressing AI bias is a core part of AI Ethics.

Related Terms