Unit 5: Ethics and Challenges of AI - Practice Quiz

1 What is bias in an AI system?

2 What can cause bias in an AI system?

3 What does AI surveillance usually involve?

4 Why can AI surveillance raise ethical concerns?

5 Which principle is important when using AI to monitor people?

6 How can AI contribute to job loss?

7 Which action can help workers affected by AI automation?

8 What is automation?

9 What is misinformation?

10 How can AI spread misinformation?

11 What should a person do before sharing AI-generated information?

12 What is a deepfake?

13 Which human right can be affected by unfair AI decisions?

14 Why should AI systems respect human rights?

15 Which right is closely connected to the use of personal data by AI?

16 What does responsible AI use mean?

17 Why is human oversight important in AI systems?

18 Which practice supports responsible AI?

19 What could happen if a biased AI is used for hiring?

20 What is one possible harm from incorrect AI medical advice?

21 A university uses an AI system to shortlist scholarship applicants. The system consistently selects fewer students from one demographic group, even though their academic qualifications are similar to those of selected students. What is the most appropriate first step?

22 A facial-recognition system performs accurately for most users but has a much higher error rate for people with darker skin tones. Which response best addresses the ethical concern?

23 An AI hiring tool uses an applicant's previous job title as a feature. Historical hiring practices favored men for senior roles, so the tool ranks male applicants higher. Why is the job title a problematic feature in this situation?

24 A city plans to use AI cameras to identify people attending public protests. Which safeguard is most important before deployment?

25 An employer uses AI to analyze workers' keystrokes and webcam images throughout the day. Which practice would make this monitoring more responsible?

26 A school wants to use an AI system to detect possible violence by analyzing students' social media posts. What is the greatest ethical risk?

27 A bank introduces AI chatbots that handle routine customer questions. Which policy best manages the effect on employees?

28 A factory adopts robots that reduce the need for manual inspection. Which conclusion is most accurate?

29 A delivery company uses an AI scheduling system that increases efficiency but gives drivers unpredictable working hours. What should the company evaluate besides productivity?

30 An AI-generated video falsely shows a public official announcing a dangerous policy. What is the most reliable response for someone who receives the video?

31 A recommendation algorithm repeatedly shows a user posts that support one political viewpoint. What risk does this create?

32 A language model produces a convincing but incorrect citation in a student report. Which practice best reduces this problem?

33 A government uses an AI system to decide which citizens receive access to essential healthcare services. Which principle requires a person to review disputed decisions?

34 A company uses an AI system to infer employees' emotional states from facial expressions and uses the results for promotions. Which human-rights concern is most relevant?

35 An AI translation service performs poorly for a minority language, limiting speakers' access to public information. Which action best supports equal rights?

36 Before deploying an AI system that approves insurance claims, a company conducts impact assessments, tests for bias, and creates an appeal process. What does this demonstrate?

37 A hospital wants to use an AI diagnostic tool. Which implementation plan is most responsible?

38 A lender's AI denies loans to applicants from a particular neighborhood because the system learned from historically unequal approval data. What harm has most directly occurred?

39 A social media platform uses AI to maximize engagement by promoting sensational false stories. What is a likely real-life consequence?

40 A police department relies heavily on an AI risk score that incorrectly labels people from one community as more likely to offend. What is the most serious consequence?

41 A hiring model selects 40% of applicants from Group A and 20% from Group B. Among qualified applicants, the selection rates are 60% for Group A and 30% for Group B. Which conclusion is most defensible?

42 A facial-recognition system has equal accuracy across demographic groups in laboratory testing but performs substantially worse for one group in public deployment. Which intervention most directly addresses the likely ethical risk?

43 An admissions model uses postal codes because they improve predictive accuracy. However, the codes strongly correlate with historical segregation. What is the strongest ethical objection?

44 A medical model has identical false-positive rates across two groups but a much higher false-negative rate for one group. Which statement is correct?

45 A city deploys predictive policing software trained on historical arrest records. Arrests are concentrated in neighborhoods that were previously over-policed. What is the most serious feedback-loop concern?

46 A workplace uses continuous AI monitoring and retains identifiable productivity data indefinitely. Which safeguard most directly applies the principle of data minimization?

47 A school installs emotion-recognition cameras to identify students who may be disengaged. Which evaluation is most ethically appropriate before deployment?

48 An automated system replaces routine tasks but creates new technical roles requiring advanced training. Which conclusion best captures the labor impact?

49 A company introduces an AI scheduling system that increases efficiency but gives unpredictable shifts to low-wage workers. Which policy best addresses the ethical problem?

50 A firm claims that AI-related layoffs are ethically neutral because affected employees receive the legally required notice. Which critique is strongest?

51 A generative model produces a plausible but false citation in a policy report. Which control most directly reduces this failure before publication?

52 A political deepfake is released shortly before an election and is disproven after voting begins. Which feature makes the harm especially difficult to reverse?

53 A recommendation algorithm maximizes engagement and repeatedly promotes sensational false claims. Which redesign best addresses the underlying incentive problem?

54 A government uses an AI system to rank citizens for access to public benefits, but affected people cannot learn the reasons for adverse decisions. Which right-related concern is most direct?

55 A content moderation model removes posts written in a minority language because its training data are limited. Which human-rights risk is most salient?

56 A humanitarian agency wants to use biometric identification to distribute aid, but refusal could prevent access to food. Which principle should guide the decision most strongly?

57 A hospital plans to deploy an AI triage tool whose training data exclude rural patients. What is the most responsible next step?

58 Which governance arrangement best preserves accountability when an AI tool assists in high-stakes loan decisions?

59 A credit model rejects applicants because it learned from historical approvals that favored homeowners, even when income and repayment indicators are comparable. What caused the harm?

60 An autonomous vehicle's vision system performs poorly in heavy rain, but the manufacturer markets it as fully autonomous without warning users. Which failure is most central?