karaza
Responsible AI

Algorithmic bias

What is algorithmic bias?

Systematic unfairness in AI output, often inherited from skewed training data or design choices, that disadvantages certain groups.

Example

A screening tool trained on a company’s past hires ranks graduates of one university higher because most past hires studied there.

Reviewed by Mahmoud, founder of Karaza, L3 in Education and TeachingUpdated