Data Ethics Debate

The data ethics debate

Computers collect huge amounts of personal data, and increasingly make decisions using AI. That raises hard questions about what is ethical (right) and what is legal (allowed). This module weighs up the issues around personal data, around AI, machine learning and robotics, and how intellectual property is protected.

Key terms

Four ideas run through the whole topic:

What does consent mean?

In data protection, what does it mean for a person to give consent?

• They have given clear permission for their data to be collected and used
• A company has sold their data on to another business
• Their data has been permanently deleted
• Their data has been scrambled so no one can read it

Match each protection to what it covers

AI: promise and problems

AI, machine learning and robotics bring real benefits, but also raise serious issues.

Issues raised by AI

Select the TWO genuine ethical or legal issues raised by AI and machine learning.

• Algorithmic bias, where the system treats some groups unfairly
• Legal liability, deciding who is responsible when AI causes harm
• AI computers run slightly faster than older machines
• AI software can be shown in any colour on screen

Be specific, not vague

In the exam, an AI answer that just says "AI can be unfair" earns little. The marks are for a specific issue tied to a real situation. For example: "If a self-driving car uses a biased algorithm and causes a crash, who is legally liable: the owner, the manufacturer, or the programmer?" Always name the issue AND link it to a scenario.

Complete the picture

The right to control who sees your data is called _____, and clear permission to use it is called _____. When an AI system treats some groups unfairly, this is called algorithmic _____. A new invention can be protected by a _____, while a song or piece of code is protected by _____.

Spot the strong points

A student is answering an AI ethics question. Tap the TWO points that are strong because each names a specific issue linked to a scenario.

• A recruitment AI trained on past hires rejects more female applicants, which is algorithmic bias.
• AI is bad and should just be banned.
• If a delivery robot injures someone, it is unclear whether the owner or the maker is liable.
• Computers can be unfair sometimes.
• AI is quite clever these days.

Plan a balanced answer

Plan a strong answer to "Discuss the ethical issues of a company using AI to screen job applicants." Put the plan in a sensible order.

• Name a specific issue: the AI could show algorithmic bias
• Link it to the scenario: it may reject certain groups unfairly
• Explain a consequence: qualified people are treated unfairly and the firm may break the law
• Give another view: AI is faster and can reduce human bias if trained well
• Reach a balanced conclusion on whether it is acceptable

You decide

Identify the main issue in each situation.

• An app shares user data with advertisers without asking permission first. What is the main issue?
• A self-driving car makes a biased decision and causes a crash. What legal question does this raise?
• Someone copies and sells a song made by another artist, without permission. What have they broken?

Make the case

A hospital wants to use an AI system to decide which patients are seen first. Discuss the ethical and legal issues, referring to a specific issue.

• Name a specific issue, such as algorithmic bias in the AI
• Link it to the scenario: how it could affect certain patients
• Explain who might be accountable or legally liable if it goes wrong
• Give a balanced view and reach a conclusion