Bias in AI and real-world consequences in AI for Everyone - Mini Project: Build & Apply

Bias in AI and Real-World Consequences

📖 Scenario: You are part of a team developing an AI system that helps decide loan approvals. You want to understand how bias in AI can affect real people and what steps you can take to identify and reduce bias.

🎯 Goal: Build a simple example that shows how bias can appear in AI data and how it can lead to unfair decisions. Learn to recognize bias and think about ways to reduce it.

📋 What You'll Learn

Create a dataset representing loan applicants with attributes including gender and income

Add a threshold value to decide loan approval

Write logic to identify biased decisions based on gender

Add a final note explaining the impact of bias on real people

💡 Why This Matters

🌍 Real World

AI systems are used in many areas like loans, hiring, and policing. Bias in data can cause unfair treatment of people based on gender, race, or other factors.

💼 Career

Understanding bias is crucial for AI developers, data scientists, and anyone working with automated decision systems to ensure ethical and fair outcomes.

1

DATA SETUP: Create the loan applicants dataset

Create a list called applicants with these exact entries as dictionaries: {'name': 'Alice', 'gender': 'female', 'income': 50000}, {'name': 'Bob', 'gender': 'male', 'income': 60000}, {'name': 'Carol', 'gender': 'female', 'income': 45000}, {'name': 'Dave', 'gender': 'male', 'income': 70000}.

AI for Everyone

# Create the list of applicants with name, gender, and income
# Your code here

Need a hint?

Use a list of dictionaries. Each dictionary should have keys 'name', 'gender', and 'income' with the exact values given.

2

CONFIGURATION: Set the income threshold for loan approval

Create a variable called income_threshold and set it to 55000. This will be the minimum income required for loan approval.

AI for Everyone

applicants = [
    {'name': 'Alice', 'gender': 'female', 'income': 50000},
    {'name': 'Bob', 'gender': 'male', 'income': 60000},
    {'name': 'Carol', 'gender': 'female', 'income': 45000},
    {'name': 'Dave', 'gender': 'male', 'income': 70000}
]
# Set the income threshold for loan approval
# Your code here

Need a hint?

Just create a variable named income_threshold and assign it the value 55000.

3

CORE LOGIC: Identify biased loan approvals

Create a list called approved that contains the names of applicants who have income greater than or equal to income_threshold. Use a for loop with variables applicant to check each applicant's income.

AI for Everyone

applicants = [
    {'name': 'Alice', 'gender': 'female', 'income': 50000},
    {'name': 'Bob', 'gender': 'male', 'income': 60000},
    {'name': 'Carol', 'gender': 'female', 'income': 45000},
    {'name': 'Dave', 'gender': 'male', 'income': 70000}
]
income_threshold = 55000
# Create a list of approved applicants based on income threshold
# Your code here

Need a hint?

Start with an empty list approved. Loop through applicants. If income is at least income_threshold, add the applicant's name to approved.

4

COMPLETION: Add a note on bias impact

Create a string variable called bias_note with this exact text: 'Notice that all approved applicants are male, showing gender bias in the data and decision process.'

AI for Everyone

applicants = [
    {'name': 'Alice', 'gender': 'female', 'income': 50000},
    {'name': 'Bob', 'gender': 'male', 'income': 60000},
    {'name': 'Carol', 'gender': 'female', 'income': 45000},
    {'name': 'Dave', 'gender': 'male', 'income': 70000}
]
income_threshold = 55000
approved = []
for applicant in applicants:
    if applicant['income'] >= income_threshold:
        approved.append(applicant['name'])
# Add a note explaining the bias impact
# Your code here

Need a hint?

Assign the exact text to a variable named bias_note.