007: bias in the machine: can an ai be unfair (ks4)
A KS4 Computer Science lesson investigating algorithmic bias and AI ethics. Students act as auditors to explore how machine learning data can lead to unfair outcomes in hiring and law.
The Mirror's Flaw: Audit an AI
The Scenario
You have been hired as a Junior Ethics Auditor for "NeuralFuture," a tech company developing a new AI for selecting job candidates. The company is worried about "Algorithmic Bias." Your manager needs a briefing document explaining the risks of using AI in hiring and examples of where this has gone wrong in the past.
The Persona: The Responsible Innovator
You are the conscience of the technical team. While developers focus on speed and efficiency, you focus on fairness, law, and society. You ask the difficult questions about how technology affects real people.
You have been hired as a Junior Ethics Auditor for "NeuralFuture," a tech company developing a new AI for selecting job candidates. The company is worried about "Algorithmic Bias." Your manager needs a briefing document explaining the risks of using AI in hiring and examples of where this has gone wrong in the past.
You are the conscience of the technical team. While developers focus on speed and efficiency, you focus on fairness, law, and society. You ask the difficult questions about how technology affects real people.
Task Instructions
1
Define the Threat
Open a new document. Title it AI Ethics Audit.
Research and explain the term Algorithmic BiasI have no idea what this means.
In your own words, explain the phrase "Garbage In, Garbage Out" in the context of Machine LearningI have no idea what this means training data.
Hint: If an AI is trained only on CVs from men, who will it prefer to hire?
2
Gather Evidence (Case Studies)
An auditor needs proof. Use a search engine to find information on ONE of the following real-world failures:
Amazon's scrapped AI recruiting tool (biased against women).
COMPAS recidivism algorithm (biased against defendants based on race).
Joy Buolamwini's research on facial recognition (Gender Shades project).
Write a summary of the case answering: What was the AI supposed to do? Who was treated unfairly? Why did the data cause this error?
3
Conduct the Audit
Imagine NeuralFuture wants to build an AI to grade student essays automatically.
Create a list of 3 potential biases this system might have. (Think about: accents in voice-to-text, handwriting styles, use of slang, or cultural references).
For each potential bias, suggest a control measureI have no idea what this means (a rule or check) to prevent it.
4
The Verdict
Conclude your report with a judgment statement answering this question: "To what extent can we trust AI to make life-changing decisions (like hiring or sentencing) without human supervision?"
Use the sentence starter: "To a large extent, AI offers efficiency, however..."
Outcome checklist
A clear definition of Algorithmic Bias and "Garbage In, Garbage Out".
A summary of a real-world case study (Amazon, COMPAS, or Facial Recognition).
An audit of the hypothetical Essay Grading AI with identified risks.
A balanced conclusion on the use of AI in high-stakes decisions.
Last modified: January 9th, 2026
