Skip to content
atlas

Don't confuse these

AI bias vs Hallucination

Why they differ

Bias is a steady tilt in results for certain groups; a hallucination is a confident answer that is simply made up.

AI bias

AI risk & governance

When an AI system treats some people or cases unfairly because of one-sided data or design choices.

Formal

A systematic error in which a model's results favour or harm certain groups, usually because the training data under-represents them or reflects past unfair decisions, or because of how the goal was defined.

In plain English

Like a new manager who only ever learned from one old boss's hiring records - they copy every habit, fair or not, and do it at speed.

In practice

A municipality tests an AI tool, trained on years of past case decisions, that ranks unemployed citizens for extra help, and finds it places people from certain areas lower although where they live was never meant to count.

Why it matters

An unfair pattern repeats in every automated decision until someone looks for it, so it can break equal-treatment and data protection law for thousands of people before anyone notices.

Hallucination

Language models

When an AI model states something false or made up - a fact, a quote, a source - in the same confident tone as a true answer.

Formal

Fluent, plausible output that is supported by neither the input nor fact, produced by an AI model that creates new text or images; it arises because such models are built to produce likely text, not checked truth.

In plain English

Like a guest at a party who never admits not knowing - asked anything, they give a smooth, detailed answer, whether or not it is true.

In practice

An official in a ministry asks a chat assistant for court rulings to support a draft reply; it names three, and a check shows that none of the cases exist.

Why it matters

Made-up answers look exactly like real ones, so any AI output used for decisions, advice or code needs a person who checks it.

Shared connections

Atlas is in beta.