AP Computer Science A: Lesson 4.1

Ethical & Social Issues Around Data Collection

Unit 4: Data Collections

This notebook serves as a complete lesson plan, presentation guide, and homework assignment for Topic 4.1.

1. Reference Guide

Key Topics

Term Definition Example / Context
Privacy & Data Ownership Who owns user data once collected? What risks exist when sensitive data is breached or sold? Data breaches exposing sensitive user records.
Algorithmic Bias Automated systems in hiring, lending, and law enforcement reinforcing historical biases. An automated screening system ranking applicants based on historic admissions data from the past 30 years.
Data Set Fitness Using incomplete, unrepresentative, or skewed datasets leads to incorrect conclusions. An AI health app predicting cardiac risk trained entirely on 20-something male professional athletes.
  • As software engineers, your algorithms are only as unbiased as the data you feed them.
  • A perfectly written loop processing flawed data will produce flawed, and often harmful, real-world decisions.

2. LxD Cycle Process

Empathize

Students often think code syntax is purely objective and neutral, assuming that because a program compiles without syntax errors, its outcomes must be fair.

Define

  • POV: CSA students need to see how "fitness" in computer science means checking if data is suitable for the purpose of the algorithm, because training a model on a subset of the population means it will fail when applied to everyone else.
  • Learning Goal: Students will analyze data collection practices, identify algorithmic bias, and propose ethical corrections to a flawed data collection pipeline.

Ideate

  • HMW Question: How might we get students to audit a dataset to ensure equity before deploying the model?
  • HMW Question: How might we show that perfectly bug-free Java code can still cause real-world harm?
  • Activity: Discuss an interactive case study about a flawed fitness app and break into pairs to analyze a university admissions algorithm.

Prototype

  • A slide deck on core concepts, an interactive case study, a Socratic discussion activity, and a student homework assignment handout.
  • Students write a 300-word reflection analyzing the root cause of algorithmic failure and detailing their developer response.

Test

  • Ask peers in peer review to discuss scenarios and trace failures back to the data collection phase.
  • Observe whether peers can identify how historical bias might be embedded in past decisions.

3. College Board Requirements

AP CSA Unit 4, Topic 4.1 Ethical and Social Issues Around Data Collection.

  • Computing innovations can reflect existing human biases.
  • Thoughtful data curation is essential for ethical software engineering.

4. Lesson Plan

Learning Objective

Analyze the relationship between data collection practices, dataset fitness, and algorithmic bias.

Success Criteria

You can identify sources of bias in a dataset, explain the concept of data set fitness, and propose ethical corrections to a flawed data collection pipeline.

Presentation & Slide Deck

Slide 1: Welcome & Warm-Up

  • Title: Data Collection: Bias, Privacy, and Social Impact
  • Warm-Up Prompt: "If a programmer writes 100% bug-free Java code, can the program still cause real-world harm? Why or why not?"
  • Teacher Script: "Welcome everyone! Today we are taking a pause from code syntax to talk about something just as critical: Data Set Fitness. As software engineers, your algorithms are only as unbiased as the data you feed them. A perfectly written loop processing flawed data will produce flawed, and often harmful, real-world decisions."

Slide 2: Core Concepts

  • Privacy & Data Ownership: Who owns user data once collected? What risks exist when sensitive data is breached or sold?
  • Algorithmic Bias: Automated systems in hiring, lending, and law enforcement reinforcing historical biases.
  • Data Set Fitness: Using incomplete, unrepresentative, or skewed datasets leads to incorrect conclusions.
  • Teacher Script: "When we talk about 'fitness' in computer science, we mean: Is this data suitable for the purpose of this algorithm? If you train a model on data that only represents a subset of the population, your algorithm will fail when applied to everyone else."

Slide 3: Interactive Case Study — "The Flawed Fitness App"

  • Scenario: A tech company launches an AI health app to predict cardiac risk. The dataset used to train the algorithm consisted entirely of 20-something male professional athletes.
  • Class Discussion Questions:
    1. What happens when this software evaluates an elderly patient or a female patient?
    2. Where did the failure occur—in the Java logic, or in the data collection phase?
    3. Who is legally and ethically responsible for this mistake?

In-Class Socratic Discussion Activity

Group Exercise (15 mins)

Break into pairs and discuss the following scenario:

A university uses an automated screening system to rank applicants. The algorithm was trained on historic admissions data from the past 30 years.

Homework Reflection Template

  1. System Background: What was the intended purpose of the software tool?
  2. Root Cause Analysis: How did flawed data collection or unrepresentative training data cause the system to fail?
  3. Developer Response: If you were the lead Java developer on this project, what specific changes would you make to the data collection process to eliminate this bias?
# Write your homework reflection below in plain text or markdown comments:
'''
Student Name: 
Date: 
Case Study Chosen: 

1. System Background:


2. Root Cause Analysis:


3. Developer Response:
'''