Ploba logo

Find the right AI tool for the job: apps, agents, APIs, MCP servers and datasets, explained plainly.

Get the AI Digest

Discover

  • Explore
  • Top Charts
  • Categories
  • Agents
  • MCP servers
  • Developer tools

Learn & community

  • Learn AI
  • AI Glossary
  • Blog
  • Hackathons
  • People
  • AI Digest

For makers

  • List your tool
  • Docs
  • Claim a listing
  • Advertise
  • Write for us

Company

  • About
  • Careers
  • Changelog
  • Trust
  • Security
  • Contact

© 2026 Ploba. All rights reserved.

  • Privacy
  • Terms
  • Cookies
  • Advertising
  • All policies
llms.txt
AI glossary
Safety

AI Bias

Also called Algorithmic bias

Systematic skew in AI outputs resulting from flawed data, design, or deployment choices.

What is AI Bias?

AI bias is a systematic skew or prejudice in the outputs of an artificial intelligence system. It occurs when an algorithm produces results that are consistently unfair, inaccurate, or discriminatory toward specific groups of people or concepts.

How does it work?

A common misconception is that AI bias comes solely from flawed training data. While biased data (e.g., a hiring dataset containing mostly male resumes) is a massive contributor, bias can also arise from:

  • Labelling: Humans providing subjective or prejudiced labels during supervised learning.
  • Model design: Developers optimizing the algorithm for metrics that inadvertently harm a demographic.
  • Evaluation: Testing the model on a benchmark that lacks diversity, giving a false sense of fairness.
  • User interaction: Users deliberately prompting the model to reinforce their own biases.

What is a simple example?

If a facial recognition system is trained primarily on images of people with lighter skin tones, the resulting model will perform significantly worse when attempting to recognize people with darker skin tones. This is a direct consequence of biased training data representation.

Why does it matter?

AI models are increasingly used to make critical decisions about loan approvals, criminal justice sentencing, and medical diagnoses. If these systems are biased, they act as scalable engines of discrimination, automating and amplifying societal inequalities.

Category
Safety
Last reviewed
Sep 21, 2026
Also called
Algorithmic bias

Sources

  • IBM: AI Bias

Related terms

  • FundamentalsTraining DataThe specific subset of a dataset used to teach an AI model how to recognize patterns and make predictions.
  • SafetyAI AlignmentThe effort to ensure an AI system’s behavior and objectives reliably match intended human goals and constraints.
  • Machine LearningModel EvaluationThe process of testing a trained AI model against a separate dataset to measure its accuracy, reliability, and safety.