Ethics and bias in AI in Intro to Computing - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Bias in AI Systems

Which of the following best describes bias in an AI system?

A AI making decisions based on incomplete or unfair data leading to unfair outcomes

B AI always producing random results regardless of input data

C AI systems that never make mistakes or errors

D AI that only works with numerical data and ignores text

💡 Hint

Think about how data quality affects AI decisions.

Bias in AI happens when the data used to train the system is incomplete or unfair, causing the AI to make unfair or incorrect decisions.

❓ trace

intermediate

Tracing Bias Impact on AI Decision

Consider an AI system trained to approve loan applications. It was trained mostly on data from one city. What is the likely result when it evaluates applications from a different city?

A It will approve all applications regardless of data

B It may unfairly reject applications from the new city due to unfamiliar data patterns

C It will reject all applications from the original city

D It will randomly approve or reject applications without any pattern

💡 Hint

Think about how training data affects AI predictions on new data.

If the AI only learned from one city's data, it might not understand the different patterns from another city, causing unfair rejections.

❓ identification

advanced

Identifying Ethical Concerns in AI Use

Which of the following scenarios raises an ethical concern about AI use?

A An AI system that recommends job candidates but only uses gender and age as criteria

B An AI that helps doctors by analyzing medical images to detect diseases

C An AI that translates languages instantly for travelers

D An AI that suggests recipes based on available ingredients

💡 Hint

Consider fairness and discrimination in AI decisions.

Using gender and age alone to recommend job candidates can lead to unfair discrimination, which is an ethical issue.

❓ Comparison

advanced

Comparing Data Sets for Fair AI Training

Which data set is best suited to train an AI system to fairly recognize faces of all people?

A A data set with images of animals instead of people

B A data set with images mostly of one ethnicity and age group

C A data set with images of people from many different ethnicities, ages, and genders

D A data set with only cartoon faces

💡 Hint

Think about diversity and representation in training data.

Diverse data helps AI learn to recognize all types of faces fairly, reducing bias.

🚀 Application

expert

Applying Ethical Principles to AI Design

You are designing an AI system for loan approvals. Which approach best reduces bias and promotes fairness?

A Train AI only on past approved loans without checking for fairness

B Use only data from one region and automate all decisions without human checks

C Ignore data quality and focus on speed of decision-making

D Use diverse training data, regularly test for bias, and allow human review of decisions

💡 Hint

Consider how to ensure fairness and accountability in AI.

Using diverse data, testing for bias, and including human review helps create fair and ethical AI systems.