The Problem of Algorithmic Bias | AI Learning Guide

Imagine an AI tasked with hiring employees. If we train it only on past hiring data from a company that mostly hired men, the AI doesn't just learn skills; it learns gender bias. It starts penalizing resumes with words like 'women's'. This is algorithmic bias: the machine amplifying historical inequalities because it treats history as truth. In the interactive simulation, you'll see how changing the data changes the fairness of the outcome.

0:00 / 0:00

Auto-play

What is Algorithmic Bias?

Algorithmic bias refers to systematic and repeatable errors in a computer system that create unfair outcomes, such as privileging one arbitrary group of users over others. Unlike random errors, bias is structural. It often stems from the training data itself, which may contain historical inequalities, or from the design choices made by developers who may unconsciously embed their own perspectives into the system's goals.

Historical Data Example

Consider a hiring algorithm trained on ten years of resumes from a male-dominated industry. The model learns that 'male' is a strong predictor of success because most successful hires were men. Consequently, it may downgrade resumes containing the word 'women's' (e.g., 'women's chess club captain'), effectively automating discrimination based on historical patterns rather than current merit.