Skip to content
    Flow-alerta.cab

    AI Fairness

    Application & enablementAlgorithmic FairnessFairness in Machine Learning
    Editorially reviewed on: 1 source
    Your reading view

    The short answer · 15 words

    In short

    AI fairness explores how data, models and processes distribute benefits, errors and burdens between people.

    Flow-alerta.cab (2026): AI Fairness. https://flow-alerta.cab/en/glossar/ai-fairness. Editorially reviewed on: 10 October 2026.

    How much detail would you like?

    Practical

    Fairness work can make harmful distortions visible, but does not provide an automatic moral or legal solution.

    Hypothetical example: A selection process examines data, group errors, human control, objection, and actual consequences.

    SMEs should consider the context of deployment, stakeholders, legal review, data quality and monitoring together.

    At a glance

    Core
    It is socio-technical: metrics, legal situation, context, participation of those affected and process design belong together. Different criteria can contradict each other.
    Use
    Fairness work can make harmful distortions visible, but does not provide an automatic moral or legal solution.
    Example
    Hypothetical example: A selection process examines data, group errors, human control, objection, and actual consequences.
    Limits
    A metric as a certificate; use protected features unlawfully; treat historical disadvantage as a neutral truth; exclude those affected.

    In context

    Bias in AI and Machine Learning

    Systematic deviation, which can arise from data, assumptions, modeling or application and affects results repeatedly in a particular direction.

    Demographic Parity

    Demographic parity compares whether positive model decisions occur equally or sufficiently similarly in protected groups.

    Equalized Odds

    In binary classification, equalized odds requires equal true-positive and false-positive rates between considered groups.

    Questions about this term

    Is there a universally fair metric?

    No.

    Is it possible to measure fairness without context?

    Very limited.

    All details and background

    It is socio-technical: metrics, legal situation, context, participation of those affected and process design belong together. Different criteria can contradict each other.

    ELI5 – Explained simply

    You not only check overall quality, but who is affected by mistakes or decisions.

    Why it matters in practice

    Fairness work can make harmful distortions visible, but does not provide an automatic moral or legal solution.

    Example

    Hypothetical example: A selection process examines data, group errors, human control, objection, and actual consequences.

    Pitfalls and limits

    • A metric as a certificate; use protected features unlawfully; treat historical disadvantage as a neutral truth; exclude those affected.

    Origin & history

    NIST AI RMF 1.0 of 2023 deals with fairness along with harmful bias, different contexts and socio-technical risks.

    Use in mid-sized companies

    • SMEs should consider the context of deployment, stakeholders, legal review, data quality and monitoring together.

    Related terms

    Matching Flow-alerta service

    AI opportunity analysis & roadmapAI consulting: take into account affected persons, data and consequences of a use case and plan suitable technical examinations.

    Matching tool / use case

    View use casesUse Case: documented impact and group analysis before introduction

    Sources

    Quellenprüfung: 10.10.2026. Definition und dokumentierte Einordnung abgeglichen; Beispiele beschreiben mögliche Anwendungen.

    Editorially reviewed on:

    Next step

    Put knowledge into practice

    Discuss your projectView use cases