Bias and AI
Start
Generative AI models are trained using large amounts of data, often sourced from the internet.
Some of these sources are reliable and fact based.
Some are from where people have shared their own ideas and personal opinions.
This data contains the answers to many questions and solutions to problems.
Unfortunately not all information on the internet is fair and balanced. It includes stereotypes and unfair ideas.
Reflecting society.
The internet also reflects society biases, discrimination and the prejudices through the content posted by its users.
As a result, AI programs may generate content based on these biases.
They may create discriminatory content based on a range of characteristics, including:
Race
Gender
Age
Health
Religion
Disability
Examples of bias in AI images
"A librarian."
"A group of people."
"A pilot."
Review before use.
Because of this, it's important to review the AI outputs to ensure we’re not spreading misinformation, discrimination, or unfair beliefs.
Great job!
You've learned about AI bias and how to protect against it.
Restart
An AI generated image of a librarian is more likely to be of an older woman.
An image of a pilot is likely to be of a white man.
A group of young people will include more people with white skin.
Bias and AI
Mary Booth
Created on May 19, 2025
Start designing with a free template
Discover more than 1500 professional designs like these:
View
Microlearning: When to Use Chat, Meetings or Email
View
Microlearning: Graphic Design
View
Microlearning: Enhance Your Wellness and Reduce Stress
View
Microlearning: Teaching Innovation with AI
View
Microlearning: Design Learning Modules
View
Word Search: Corporate Culture
View
Professional quiz mobile
Explore all templates
Transcript
Bias and AI
Start
Generative AI models are trained using large amounts of data, often sourced from the internet.
Some of these sources are reliable and fact based.
Some are from where people have shared their own ideas and personal opinions.
This data contains the answers to many questions and solutions to problems.
Unfortunately not all information on the internet is fair and balanced. It includes stereotypes and unfair ideas.
Reflecting society.
The internet also reflects society biases, discrimination and the prejudices through the content posted by its users.
As a result, AI programs may generate content based on these biases.
They may create discriminatory content based on a range of characteristics, including:
Race
Gender
Age
Health
Religion
Disability
Examples of bias in AI images
"A librarian."
"A group of people."
"A pilot."
Review before use.
Because of this, it's important to review the AI outputs to ensure we’re not spreading misinformation, discrimination, or unfair beliefs.
Great job!
You've learned about AI bias and how to protect against it.
Restart
An AI generated image of a librarian is more likely to be of an older woman.
An image of a pilot is likely to be of a white man.
A group of young people will include more people with white skin.