Unit 4: A4 Machine learning
Topic 4.4: Ethics and impact Questions
Practice 20 exam-style questions for IB Computer Science Topic 4.4. Review the question stems below, then unlock the full Question Bank to access markschemes, model answers, and AI grading.
1Choose1 mark
2026
Choose the most likely source of bias in a machine learning system.
Markscheme and model answer locked
Unlock Question2Choose1 mark
2026
Choose why removing a sensitive attribute from the training data may not remove bias.
Markscheme and model answer locked
Unlock Question3Choose1 mark
2026
Choose the principle that says data should be used only for the purpose it was collected for.
Markscheme and model answer locked
Unlock Question4Suggest4 marks
2026
Suggest who should be responsible when an automated system causes harm.
Markscheme and model answer locked
Unlock Question5Explain4 marks
2026
Explain why a person refused a loan by an automated system may have difficulty challenging the decision.
Markscheme and model answer locked
Unlock Question6Suggest4 marks
2026
Suggest what an organisation should provide alongside an automated decision.
Markscheme and model answer locked
Unlock Question7Determine3 marks
2026
Determine which error matters more for a model screening medical images, and justify your answer.
Markscheme and model answer locked
Unlock Question8Suggest3 marks
2026
Suggest why the right to have personal data deleted is difficult to honour for a trained model.
Markscheme and model answer locked
Unlock Question10State1 mark
2026
State one source of bias in a machine learning system.
Markscheme and model answer locked
Unlock Question11Explain4 marks
2026
Explain how a recruitment model trained on a company's past hiring decisions could become biased.
Markscheme and model answer locked
Unlock Question12Suggest4 marks
2026
Suggest what would make human review of automated decisions meaningful.
Markscheme and model answer locked
Unlock Question13Explain3 marks
2026
Explain how a facial recognition system could work less well for some groups than others.
Markscheme and model answer locked
Unlock Question14Suggest4 marks
2026
Suggest two ways of detecting bias in a deployed model.
Markscheme and model answer locked
Unlock Question15Suggest4 marks
2026
Suggest two ways of reducing bias when a model is built.
Markscheme and model answer locked
Unlock Question16Explain4 marks
2026
Explain the concern raised by a model trained on personal data scraped from the internet.
Markscheme and model answer locked
Unlock Question17Explain4 marks
2026
Explain why a model that is accurate overall may still be unacceptable.
Markscheme and model answer locked
Unlock Question18Explain4 marks
2026
Explain why automating a decision changes its ethical weight even when the decision itself is unchanged.
Markscheme and model answer locked
Unlock Question19Explain4 marks
2026
Explain the concern about machine learning systems replacing human judgement in criminal justice.
Markscheme and model answer locked
Unlock Question20Suggest3 marks
2026
Suggest one legitimate use of machine learning in a high-stakes setting.
Markscheme and model answer locked
Unlock QuestionReady to practice Topic 4.4?
Get instant AI feedback on your answers, view detailed markschemes, and track your progress across all IB Computer Science topics.
Start Practising Now