Wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

AI Ethics and Responsible AI Quiz

Total questions: 85

Worksheet time: 43mins

Name
Class
Date
1.

An eco-friendly food distribution center is using AI to reduce waste. They ensure their AI models are used to benefit people and society while avoiding harm. What does this scenario describe?

a)

Responsible AI

b)

Dynamic AI

c)

Progressive AI

d)

Regenerative AI

e)

None of the Above

2.

A bank's AI model incorrectly and unfairly denies loans to applicants from a specific neighborhood, even though their qualifications are the same as those of approved applicants from other areas. What type of harm does this scenario describe?

a)

Knowledge cutoff

b)

Representational

c)

Allocative

d)

Deepfakes

e)

None of the Above

3.

An AI model used for predicting real estate trends has started to decline in accuracy. Which of the following factors could be contributing to the drift in this AI model? Select all that apply.

a)

A change in economic policies affecting the housing market

b)

A stable pattern of seasonal home sales

c)

New real estate regulations

d)

A sudden shift in buyer preferences

e)

None of the Above

4.

Which two statements best describe data bias? Select two answers.

a)

Data bias can exist within high-quality training data

b)

Data bias is a circumstance in which systemic errors lead to inaccurate information

c)

Data bias ensures that AI models are trained on the latest information available

d)

Data bias is a feature that enables AI models to fix errors in their programming

e)

None of the Above

5.

An accountant is using AI-powered accounting software to help complete tasks for their clients. What steps should the accountant take while using this software to ensure their clients' data remains private and secure? Select all that apply.

a)

Create thorough inputs by including many specific and personalized details

b)

Familiarize themselves with new advancements in AI and their related risks

c)

Understand the AI tool’s data-collection practices

d)

Read the AI tool’s terms of use or service

e)

None of the Above

6.

A company uses AI to recommend products to customers but notices that recommendations favor one demographic over others. What type of bias is likely occurring?

a)

Representation bias

b)

Allocation bias

c)

Sampling bias

d)

Confirmation bias

e)

None of the Above

7.

A healthcare AI tool incorrectly predicts patient risks for a particular age group. How should the developers address this issue?

a)

Ignore the discrepancy

b)

Retrain the model with diverse patient data

c)

Use the AI for only other age groups

d)

Reduce the dataset size

e)

None of the Above

8.

An AI chatbot spreads misinformation accidentally due to outdated knowledge. Which type of harm does this illustrate?

a)

Knowledge cutoff

b)

Representational

c)

Allocative

d)

Deepfakes

e)

None of the Above

9.

An AI system for hiring resumes applicants only from specific universities, excluding equally qualified candidates elsewhere. What is the primary concern?

a)

Allocative harm

b)

Representational harm

c)

Knowledge cutoff

d)

Data drift

e)

None of the Above

10.

A company wants to ensure that its AI recommendations do not unintentionally harm users. Which principle are they following?

a)

Responsible AI

b)

Dynamic AI

c)

Progressive AI

d)

Regenerative AI

e)

None of the Above

11.

An AI system misclassifies customers into wrong loyalty tiers, affecting discounts. Which type of harm is illustrated?

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

12.

A financial AI system fails to update its model after new market regulations. This is an example of:

a)

Knowledge cutoff

b)

Representational harm

c)

Allocative harm

d)

Data bias

e)

None of the Above

13.

Which of the following steps help reduce data bias in AI models? Select all that apply.

a)

Collect diverse and representative datasets

b)

Regularly audit model predictions

c)

Ignore unusual outcomes

d)

Apply bias mitigation techniques

e)

None of the Above

14.

An AI system generates content that falsely impersonates a person. This scenario illustrates:

a)

Deepfakes

b)

Representational harm

c)

Allocative harm

d)

Knowledge cutoff

e)

None of the Above

15.

Which of the following best describes “Responsible AI”?

a)

AI that is transparent, fair, and accountable

b)

AI that is fast and efficient

c)

AI that is fully autonomous without oversight

d)

AI that prioritizes profit over ethics

e)

None of the Above

16.

An AI-powered translation tool mistranslates texts from minority languages consistently. What type of bias is occurring?

a)

Representation bias

b)

Allocative bias

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

17.

A company wants to maintain user privacy while using AI. Which is an effective step? Select all that apply.

a)

Use anonymized data for AI training

b)

Limit data collection to necessary information

c)

Share personal data with third parties without consent

d)

Regularly review AI data practices

e)

None of the Above

18.

A predictive AI system recommends loans, but its recommendations systematically favor one neighborhood over another. The issue is:

a)

Allocative harm

b)

Representational harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

19.

An AI virtual assistant accidentally discloses private conversations due to a software bug. The company should:

a)

Ignore the incident

b)

Inform affected users and fix the bug

c)

Blame the users

d)

Continue using the AI without changes

e)

None of the Above

20.

Which practice helps ensure AI fairness? Select all that apply.

a)

Conduct regular audits of AI outcomes

b)

Train models on biased datasets

c)

Implement bias mitigation techniques

d)

Ignore minority group performance

e)

None of the Above

21.

An AI system used in a hospital misdiagnoses patients from a specific demographic group. Which type of harm is this?

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

22.

A self-driving car AI struggles in unusual weather conditions because it was trained mostly on sunny-day data. What is this an example of?

a)

Representation bias

b)

Knowledge cutoff

c)

Allocative harm

d)

Dynamic AI

e)

None of the Above

23.

Which of the following are ways to promote Responsible AI? Select all that apply.

a)

Conducting transparency audits

b)

Ensuring fairness in model predictions

c)

Ignoring feedback from affected users

d)

Applying accountability measures

e)

None of the Above

24.

An AI tool misclassifies resumes from minority applicants while accurately classifying others. The bias is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

25.

A company collects biometric data for AI access control. Employees must consent before data collection. This demonstrates:

a)

Ignoring user privacy

b)

Responsible AI practices

c)

Allocative harm

d)

Deepfakes

e)

None of the Above

26.

An AI-powered content moderation system flags certain dialects more than others, unfairly limiting speech. This is:

a)

Allocative harm

b)

Representational harm

c)

Knowledge cutoff

d)

Data drift

e)

None of the Above

27.

Which of the following can contribute to AI model drift? Select all that apply.

a)

Changes in market trends

b)

Shifts in consumer behavior

c)

Stable and predictable data

d)

Policy or regulatory changes

e)

None of the Above

28.

A recommendation AI consistently favors customers from wealthy neighborhoods. The harm type is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

29.

A hospital AI fails to update its training data and makes outdated medical recommendations. What principle is violated?

a)

Responsible AI

b)

Knowledge cutoff

c)

Allocative harm

d)

Representational harm

e)

None of the Above

30.

An AI tool for grading essays systematically under-scores non-native English speakers. The bias type is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

31.

er-scores non-native English speakers. The bias type is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

32.

Which practices ensure AI privacy and security? Select all that apply.

a)

Encrypting user data

b)

Sharing personal data without consent

c)

Limiting data collection

d)

Regularly reviewing AI data practices

e)

None of the Above

33.

An AI model for credit scoring unfairly favors applicants from urban areas. The type of harm is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

34.

A social media AI recommends misleading health information. What type of harm is this?

a)

Deepfakes

b)

Allocative

c)

Knowledge cutoff

d)

Representational

e)

None of the Above

35.

Which of the following help reduce bias in AI models? Select all that apply.

a)

Use representative datasets

b)

Audit AI predictions

c)

Ignore unusual results

d)

Apply bias mitigation techniques

e)

None of the Above

36.

A chatbot spreads misinformation due to outdated information. This is an example of:

a)

Knowledge cutoff

b)

Allocative harm

c)

Representational harm

d)

Deepfakes

e)

None of the Above

37.

An AI recruitment tool systematically excludes candidates from rural areas. The type of harm is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

38.

Which practices reflect Responsible AI? Select all that apply.

a)

Transparency in data use

b)

Fairness in predictions

c)

Accountability for outcomes

d)

Ignoring stakeholder concerns

e)

None of the Above

39.

An AI model fails to recommend loans fairly, disproportionately rejecting applicants from a certain neighborhood. This demonstrates:

a)

Allocative harm

b)

Representational harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

40.

A healthcare AI system provides outdated treatment suggestions because it was trained on old data. This illustrates:

a)

Knowledge cutoff

b)

Allocative harm

c)

Representational harm

d)

Deepfakes

e)

None of the Above

41.

Which of the following are steps to maintain privacy while using AI? Select all that apply.

a)

Use anonymized datasets

b)

Limit data collection to necessary information

c)

Share sensitive data without consent

d)

Conduct regular audits of data practices

e)

None of the Above

42.

An AI-powered translation tool consistently mistranslates minority languages. What type of bias is occurring?

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

43.

A company implements AI to monitor employee productivity but collects more data than necessary. What principle is violated?

a)

Responsible AI

b)

Knowledge cutoff

c)

Allocative harm

d)

Data bias

e)

None of the Above

44.

Which of the following steps help ensure AI fairness? Select all that apply.

a)

Conduct regular audits

b)

Train models on biased datasets

c)

Apply bias mitigation techniques

d)

Ignore minority group performance

e)

None of the Above

45.

An AI system for predictive policing disproportionately targets certain neighborhoods. The harm is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

46.

A virtual assistant accidentally leaks private messages due to a software bug. The company should:

a)

Inform affected users and fix the bug

b)

Ignore the incident

c)

Blame the users

d)

Continue using AI without changes

e)

None of the Above

47.

An AI tool in healthcare recommends treatments based on outdated research. What type of harm is this?

a)

Knowledge cutoff

b)

Allocative harm

c)

Representational harm

d)

Deepfakes

e)

None of the Above

48.

Which of the following are ways to reduce data bias in AI? Select all that apply.

a)

Collect diverse and representative datasets

b)

Regularly audit model predictions

c)

Ignore unexpected outcomes

d)

Apply bias mitigation techniques

e)

None of the Above

49.

An AI content moderation system flags minority dialects disproportionately. The bias is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

50.

A recommendation AI consistently favors customers from wealthy neighborhoods. The type of harm is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

51.

An AI system misclassifies applicants from a particular gender during hiring. The bias is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

52.

A company wants to ensure AI transparency and fairness. Which practices support this? Select all that apply.

a)

Conduct audits of AI outputs

b)

Use diverse training data

c)

Ignore stakeholder feedback

d)

Apply bias mitigation strategies

e)

None of the Above

53.

An AI system denies loans disproportionately to low-income applicants. The type of harm is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

54.

An AI-powered health app gives wrong advice due to outdated information. What principle is violated?

a)

Responsible AI

b)

Allocative harm

c)

Knowledge cutoff

d)

Representational harm

e)

None of the Above

55.

Which of the following are effective steps to maintain AI privacy? Select all that apply.

a)

Use anonymized datasets

b)

Limit data collection

c)

Share sensitive data without consent

d)

Conduct regular audits

e)

None of the Above

56.

An AI system consistently misclassifies minority customers’ profiles. The bias is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

57.

A financial AI model fails to update after new regulations, producing outdated recommendations. The principle violated is:

a)

Responsible AI

b)

Knowledge cutoff

c)

Allocative harm

d)

Representational harm

e)

None of the Above

58.

Which of the following support Responsible AI? Select all that apply.

a)

Transparency in data usage

b)

Fairness in predictions

c)

Accountability for outcomes

d)

Ignoring stakeholder concerns

e)

None of the Above

59.

A chatbot spreads misinformation due to outdated knowledge. The type of harm is:

a)

Knowledge cutoff

b)

Allocative harm

c)

Representational harm

d)

Deepfakes

e)

None of the Above

60.

An AI recruitment tool favors candidates from certain schools while ignoring equally qualified others. The bias is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

61.

Which of the following practices ensure AI fairness and security? Select all that apply.

a)

Encrypt sensitive data

b)

Conduct audits of AI predictions

c)

Limit collection to necessary data

d)

Ignore biased outcomes

e)

None of the Above

62.

An AI virtual assistant accidentally shares private messages due to a software bug. The company should:

a)

Ignore the incident

b)

Inform affected users and fix the bug

c)

Blame the users

d)

Continue using AI without changes

e)

None of the Above

63.

An AI system for credit scoring disproportionately rejects applicants from a certain neighborhood. The type of harm is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

64.

A healthcare AI tool provides outdated recommendations due to old training data. The principle violated is:

a)

Responsible AI

b)

Knowledge cutoff

c)

Allocative harm

d)

Representational harm

e)

None of the Above

65.

Which of the following help maintain privacy in AI systems? Select all that apply.

a)

Use anonymized datasets

b)

Limit data collection to necessary information

c)

Share sensitive data without consent

d)

Conduct regular audits of AI data practices

e)

None of the Above

66.

An AI content moderation tool flags dialects disproportionately. This is an example of:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

67.

A predictive AI for lending systematically favors applicants from certain areas. This type of harm is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

68.

Which practices promote Responsible AI? Select all that apply.

a)

Transparency in data use

b)

Fairness in model predictions

c)

Accountability for outcomes

d)

Ignoring stakeholder concerns

e)

None of the Above

69.

An AI model for grading essays under-scores non-native English speakers. The bias type is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

70.

An AI-powered health app provides inaccurate advice due to outdated knowledge. This illustrates:

a)

Knowledge cutoff

b)

Allocative harm

c)

Representational harm

d)

Deepfakes

e)

None of the Above

71.

Which of the following reduce bias in AI systems? Select all that apply.

a)

Use diverse and representative datasets

b)

Regularly audit model predictions

c)

Ignore unusual outcomes

d)

Apply bias mitigation techniques

e)

None of the Above

72.

A recruitment AI tool favors candidates from certain schools while ignoring equally qualified applicants. The bias type is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

73.

An AI model misclassifies customer profiles, disadvantaging minority groups. The type of bias is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

74.

A social media AI recommends misleading health information. This type of harm is:

a)

Deepfakes

b)

Allocative harm

c)

Knowledge cutoff

d)

Representational harm

e)

None of the Above

75.

An AI model fails to update after policy changes, producing outdated outputs. This demonstrates:

a)

Knowledge cutoff

b)

Allocative harm

c)

Representational harm

d)

Deepfakes

e)

None of the Above

76.

Which practices help ensure fairness in AI? Select all that apply.

a)

Conduct audits of AI predictions

b)

Apply bias mitigation techniques

c)

Ignore unusual results

d)

Use representative datasets

e)

None of the Above

77.

An AI tool in a hospital systematically underestimates risk for a certain demographic. The harm type is:

a)

Representation bias

b)

Allocative harm

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

78.

An AI recommendation system favors wealthy customers. This type of harm is:

a)

Allocative

b)

Representation bias

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

79.

An AI virtual assistant leaks private information due to insufficient security measures. The company should:

a)

Inform affected users and fix the issue

b)

Ignore the incident

c)

Blame the users

d)

Continue using AI without changes

e)

None of the Above

80.

A financial AI model systematically denies loans to low-income applicants. The type of harm is:

a)

Allocative

b)

Representation bias

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

81.

Which steps help maintain Responsible AI and privacy? Select all that apply.

a)

Use anonymized datasets

b)

Limit data collection

c)

Conduct regular audits

d)

Share personal data without consent

e)

None of the Above

82.

An AI virtual assistant accidentally shares private messages due to a software bug. The company should:

a)

Ignore the incident

b)

Inform affected users and fix the bug

c)

Blame the users

d)

Continue using AI without changes

e)

None of the Above

83.

An AI system for credit scoring disproportionately rejects applicants from a certain neighborhood. The type of harm is:

a)

Allocative

b)

Representational

c)

Knowledge cutoff

d)

Deepfakes

e)

None of the Above

84.

A healthcare AI tool provides outdated recommendations due to old training data. The principle violated is:

a)

Responsible AI

b)

Knowledge cutoff

c)

Allocative harm

d)

Representational harm

e)

None of the Above

85.

Which of the following help maintain privacy in AI systems? Select all that apply.

a)

Use anonymized datasets

b)

Limit data collection to necessary information

c)

Share sensitive data without consent

d)

Conduct regular audits of AI data practices

e)

None of the Above

Similar Resources on Wayground

BA 5

82 questions

BA 5

quiz

Worksheet

•

University