Speaking Boost · C1 · Lesson 26 of 50
Speaking Boost C1 · Block 6: Ethics and Technology

Should a machine make moral decisions?

Self-driving cars and automated systems already make choices that affect lives. Discuss who is responsible when things go wrong, and learn modal perfects: should have, could have, might have, must have, can't have and needn't have.

1

Warm-up

5 min
1.1

Machines and moral choices

Answer each question with 2–3 sentences.

  1. Would you ride in a car with no driver?
  2. Should a machine ever decide who gets medical treatment?
  3. Who should be blamed when an automated system makes a mistake?
  4. Do you trust software more than people?
  5. Have you ever blamed a machine for a human mistake?
  6. What is a decision that only a human should make?
2

Words and phrases

12 min
2.1

New words

Read each word and the example. Then say a sentence about machines and ethics.

2.2

Which word?

Choose the word that fits.

  1. A situation in which you must choose between two difficult options is a .
  2. The company may face legal if its software harms someone.
  3. A vehicle that can drive itself is called .
  4. A approach judges actions by whether they produce the greatest good for the greatest number.
  5. Hospitals must patients when there are too few beds, treating the most urgent first.
  6. After the accident there was one and several people with minor injuries.
  7. Without human , an automated system may repeat errors for years.
  8. A such as an emergency stop button protects against harm.
2.3

Opposites and pairs

Choose the right answer, then press Check answers.

  1. a dilemma is a choice between two
  2. liability = legal
    for harm
  3. autonomous = able to act without human
  4. to prioritise = to treat as most
  5. a safeguard protects against
  6. oversight = careful
2.4

Useful phrases

Use these phrases to judge past decisions and speculate about what happened.

  • It should have...Це мало б...It should have stopped sooner.
  • It could have...Це могло б...It could have saved more people.
  • They might have...Вони, можливо...They might have misjudged the risk.
  • It must have been...Це, мабуть, було...It must have been a software error.
  • It can't have been...Це не могло бути...It can't have been the driver's fault.
  • You needn't have...Тобі не треба було...You needn't have worried.
  • They ought to have...Вони мали б...They ought to have tested it first.
  • Who is to blame?Хто винен?Who is to blame when a machine fails?
3

Read and react

8 min
3.1

Machines and moral choices

Read the text. Hover over the dotted words for help.

Should a machine make moral decisions?

Should a machine make moral decisions? The question is no longer only theoretical. Self-driving cars, hospital software and automated weapons systems already take decisions with serious consequences, and engineers must anticipate situations that no human has described to them.

Philosophers often use the "trolley problem", introduced by Philippa Foot in 1967, to explore such choices: should you divert a runaway trolley to kill one person instead of five? A utilitarian may say yes, while others argue that deliberately causing a death is different from allowing one. In 2018, researchers published the results of the Moral Machine experiment, in which millions of people from over two hundred countries and territories judged similar dilemmas for self-driving cars. Preferences varied between regions, which suggests that a single global rule would be hard to design.

Real accidents are more complicated. In 2018, a pedestrian was killed in Arizona by a test vehicle that was driving itself, and investigators said that both technical problems and the supervision of the backup driver had played a part. Observers asked who should have prevented the crash: the engineers who could have tested the system more, the company that might have been under pressure to launch, or the driver who must have been distracted.

Many experts argue that machines should support, not replace, human judgement, and that oversight and liability must be clear. A machine cannot be punished or feel regret, so someone else must answer for its choices.

So should a machine make moral decisions? In narrow cases, perhaps, provided that humans set the rules and remain responsible.

Which decision would you never give to a machine?

3.2

True or false?

  1. The trolley problem was introduced by Philippa Foot.
  2. The Moral Machine experiment found identical preferences in every region.
  3. Investigators said both technical problems and supervision played a part in the 2018 crash.
  4. The text says that a machine can be punished.
  5. The text suggests that humans should remain responsible.
3.3

What do you think?

  1. Who should be blamed when an autonomous car crashes?
  2. Should cars be programmed to protect passengers first?
  3. Can machines have moral responsibility?
  4. Which rules should be the same in every country?
4

Grammar for speaking

10 min
4.1

Grammar for speaking

Modal perfects

Use modal + have + past participle to judge or speculate about the past. should have / ought to have = a criticism or regret; could have = a possibility that did not happen; might have = a possibility that is uncertain; must have = a strong deduction; can't / couldn't have = a strong negative deduction; needn't have = an action that was unnecessary.

The car should have stopped sooner. It could have saved more people. It must have been a software error. It can't have been the driver's fault. You needn't have worried.

Passives and continuous forms

For passive meaning use modal + have been + past participle. For an action in progress use modal + have been + -ing.

The system should have been tested before launch. They might have been driving too fast. The error could have been avoided.

Block 6 preview

Grammar you will use in this block.

modal perfects (L26) · future forms (L27) · inversion in conditionals (L28) · future in the past (L29) · review and obligation and permission (L30)

Common mistakes: modal perfects are modal + have + past participle, and have never changes to had. Use should have been + past participle for passives. ✗ The car should had stopped → should have stopped. ✗ They must knew → must have known. ✗ It can't been the driver → can't have been. ✗ You needn't to have worried → needn't have worried.

4.2

Choose the right form

  1. The car
    earlier.
  2. The driver
    distracted.
  3. They
    it again.
  4. It
    the software.
  5. The engineers
    the risks.
  6. The system
    before launch.
  7. She
    more people.
  8. He
    us.
4.3

Fix the mistake

Each sentence has one mistake. Type the correct words.

  1. The car should had stopped earlier. →
  2. They must knew about the risk. →
  3. The crash can't been the driver's fault. →
  4. You needn't to have worried. →
  5. The program should tested before launch. →
  6. She could saved more lives. →
4.4

Talk about it

Answer in detail. Use modal perfects.

  1. Think of a famous accident. What should have been done differently?
  2. What could the engineers have done to prevent a crash?
  3. What might have happened if a human had been driving?
  4. It must have been hard to decide. Why?
  5. What needn't have happened if the system had been tested?
  6. Who ought to have taken responsibility?
5

Three views and a short talk

8 min
5.1

Three views on machine ethics

Three people discuss machines and moral choices. Compare them. Who trusts machines most? Who trusts them least? Whose view is closest to yours?

Person 1

Dr Moreau, 49

  • Is an AI ethicist
  • Says humans must stay in charge
  • Warns that rules should have been set earlier
  • Believes in clear liability
Person 2

Zack, 31

  • Is a software engineer
  • Says engineers can't predict every case
  • Thinks testing could have been better
  • Wants clear standards
Person 3

Mrs Lin, 64

  • Is a retired nurse
  • Says machines can't care
  • Fears that hospitals will rely on software
  • Insists that a person must decide
5.2

Mini-talk: machines and moral choices

Your teacher will time you.

  1. Plan for one minute, then speak for one or two minutes.
  2. Part 1: describe a decision that a machine might make.
  3. Part 2: say who should be responsible if it goes wrong (use at least three modal perfects).
  4. Part 3: say what should be decided by humans only.
2:00
6

Personal questions

10 min
6.1

Part I: Me

  1. Would you trust a self-driving car?
  2. Do you use apps that make decisions for you?
  3. Have you blamed a machine for a human mistake?
  4. What decisions do you leave to software?
  5. Would you accept an automated medical diagnosis?
  6. Do you read about technology news?
  7. Do you think machines can be fair?
  8. What would worry you most about autonomous machines?
6.2

Part II: Other people

  1. Who would you blame if a robot made a mistake?
  2. Who should write the rules for machines?
  3. Who in your life works with technology?
  4. Who do you trust to check software?
  5. Who pays when an autonomous car crashes?
  6. Who taught you to think about ethics?
6.3

Part III: A bit deeper

  1. Should machines make life-and-death decisions?
  2. Can a machine be moral?
  3. Is human oversight enough?
  4. Should some technologies be banned?
  5. Are we too quick to trust machines?
7

Roleplay

12 min
7.1

Investigating an accident

You

Investigator

You investigate an accident involving an automated system (your teacher):
  • ask what happened
  • ask who could have prevented it
  • ask what should have been tested
  • give your conclusion
Speak for at least 2 minutes in total. Use modal perfects.
Your teacher

Engineer

Says things like:
  1. The car did not stop.
  2. We might have missed a case.
  3. We should have tested more.
  4. Thank you for your report.
7.2

Debating machine ethics

You

Speaker

You are in a debate (your teacher):
  • state your view on machines and moral choices
  • give an example
  • respond to an objection
  • use should have and could have
You do most of the talking. Use: It should have... · It might have... · Who is to blame?
Your teacher

Opponent

Says things like:
  1. Machines are more objective.
  2. Can you give an example?
  3. But humans make errors too.
  4. So who is responsible?
8

Homework

at home

Your homework score

Do all four exercises, then press Check all homework. Correct answers will appear next to your mistakes.

– / 30
H1

Complete with a word

dilemmaliabilityautonomousutilitarianprioritisecasualtyoversightsafeguard
  1. A situation in which you must choose between two difficult options is a .
  2. The company may face legal if its software harms someone.
  3. A vehicle that can drive itself is called .
  4. A approach judges actions by whether they produce the greatest good for the greatest number.
  5. Hospitals must patients when there are too few beds, treating the most urgent first.
  6. After the accident there was one and several people with minor injuries.
  7. Without human , an automated system may repeat errors for years.
  8. A such as an emergency stop button protects against harm.
H2

Choose the correct form

  1. The pilot earlier.
  2. The system last night.
  3. You so early.
  4. It his fault.
  5. They the problem.
  6. The data twice.
  7. We the contract.
  8. She first.
H3

Correct or incorrect?

Decide if each sentence is grammatically correct.

  1. The car should have stopped earlier.
  2. The car should had stopped earlier.
  3. They must have known about the risk.
  4. They must knew about the risk.
  5. It can't have been the driver's fault.
  6. It can't been the driver's fault.
  7. You needn't have worried.
H4

Match the two halves

  1. It should have
  2. It could have
  3. They might have
  4. It must have been
  5. It can't have been
  6. You needn't have
  7. They ought to have

Finished the lesson?

Mark it as complete to keep track of your progress.

© Yevheniia Kolesnykova, Academy of Languages. All rights reserved. No copying or reuse without written permission.