Computer Sciencecalendar_todayLast updated: Apr 2026

What is Algorithmic Bias?

/ˌælɡəˈrɪðmɪk ˈbaɪəs/

When an AI system reflects the biases of its creators or training data, producing outputs that systematically favour or disadvantage certain groups.
lightbulb

Everyday Example

A facial recognition system trained mostly on lighter-skinned faces may struggle to accurately identify darker-skinned individuals — not because of intentional prejudice, but because of imbalanced training data.

publicReal-World Application

Hiring algorithms that learn from past successful hires can inadvertently encode historical biases — for example, penalising CV keywords more common in women's applications because past hires were predominantly male.
psychology

Did you know?

The study of algorithmic bias gained mainstream attention after a 2018 MIT Media Lab study found commercial AI systems misidentified the gender of dark-skinned women at rates up to 34% higher than lighter-skinned men.

emoji_objects

Key Insight

Bias in, bias out. The quality and diversity of training data is one of the most important — and often overlooked — factors in building fair AI systems.

Want to learn Algorithmic Bias in 60 seconds?

Join 50,000+ learners snacking on knowledge daily.