What is Algorithmic Bias?

S8-SA4-0040

Grade Level:

Class 5

AI/ML, Data Science, Research, Journalism, Law, any domain requiring critical thinking

Definition

What is it?

Algorithmic bias happens when a computer program, called an algorithm, makes unfair or incorrect decisions because of the data it was trained on or how it was designed. It's like a calculator giving wrong answers for some numbers, not because it's broken, but because it learned from a faulty table.

Simple Example

Quick Example

Imagine an app that recommends songs. If this app was mostly trained on songs popular in Mumbai, it might not suggest many folk songs from Rajasthan or Punjabi pop, even if you like them. This is a bias because it's not fair to all kinds of music or listeners.

Worked Example

Step-by-Step

Let's say a school's computer system helps decide which students get a special scholarship for 'Sports Excellence'.

Step 1: The system is given past data of students who received scholarships. This data mostly includes boys who played cricket or football.

Step 2: A girl who is an excellent Kabbadi player applies. The algorithm looks at her sport.

Step 3: Because Kabbadi for girls wasn't common in the old data, the algorithm might not 'recognize' it as a top 'Sports Excellence' activity.

Step 4: The system might then rank her lower than a boy with similar skill in cricket, even though she is equally deserving.

Step 5: This leads to the Kabbadi player not getting the scholarship, not because she isn't good, but because the algorithm had a bias from the old data.

Answer: The algorithm showed bias against sports not well-represented in its training data.

Why It Matters

Understanding algorithmic bias is super important because these programs are used everywhere – from helping doctors decide treatments to suggesting news articles you read. Future journalists, scientists, and lawyers need to spot and fix these biases to ensure fairness for everyone. It helps create a more just and equal world.

Common Mistakes

MISTAKE: Thinking algorithms are always fair because they are machines. | CORRECTION: Algorithms are created by humans and learn from human data, so they can inherit human biases and mistakes.

MISTAKE: Believing bias only means a program is 'broken'. | CORRECTION: Algorithmic bias often means the program is working exactly as designed, but its design or the data it learned from was flawed or incomplete.

MISTAKE: Assuming bias only affects big, complex things like self-driving cars. | CORRECTION: Bias can appear in everyday apps too, like job recommendation systems, social media feeds, or even search results.

Practice Questions

Try It Yourself

QUESTION: An online shopping app suggests only expensive brands to you, even though you mostly buy budget-friendly items. What kind of problem might this be? | ANSWER: Algorithmic bias, possibly from data showing only high-spending customers in a similar profile.

QUESTION: A news app shows you articles only about Bollywood movies, even though you sometimes read about science and history. Why might this happen? | ANSWER: The app's algorithm might have learned from your past clicks that you prefer entertainment, creating a bias and not showing you other types of news you might like.

QUESTION: A facial recognition system struggles to identify people with darker skin tones because it was mostly trained on images of people with lighter skin. Explain how this is an example of algorithmic bias and its source. | ANSWER: This is algorithmic bias because the system performs unfairly for a specific group. The source of this bias is the 'training data' – the images used to teach the system were not diverse enough, leading it to be less accurate for darker skin tones.

MCQ

Quick Quiz

What is the main reason an algorithm might show bias?

The computer gets too hot

The internet connection is slow

The data it learned from was unfair or incomplete

The algorithm is too fast

The Correct Answer Is:

Algorithmic bias primarily comes from the data used to train the algorithm or the way it was designed, not from technical issues like heat, internet speed, or processing speed.

Real World Connection

In the Real World

Many social media apps in India use algorithms to decide which posts, ads, or videos you see. If an algorithm notices you click more on posts from your city, it might start showing you fewer posts from other states, even if they are important. This shapes your view of the world and can create a 'filter bubble' where you only see certain types of content.

Key Vocabulary

Key Terms

ALGORITHM: A set of rules or instructions a computer follows to solve a problem | DATA: Information collected, like numbers, facts, or images, that algorithms learn from | BIAS: A preference or prejudice that makes something unfair | TRAINING DATA: The specific data used to teach an algorithm how to perform a task

What's Next

What to Learn Next

Next, you can learn about 'Fairness in AI'. This concept builds on algorithmic bias by exploring how we can design and improve algorithms to make sure they are fair and do not show bias. It's about finding solutions to the problems we just discussed!