Đề Thi FE ITE302c - SP26 - FE - RE

[adminadmin is verified member.]

Học kỳ

SP2026

Thời Gian

6/5/26

Loại tài liệu

FE

Mã Đề

ITE302c_SP26_RE_605832

ITE302c SP26 FE RE

1. (Choose 1 answer)

An example of a public dataset at risk of an algorithmic privacy violation is the
A. Google Maps satellite view dataset
B. Google search view dataset
C. Stanford hospital dataset
D. Fidelity Bank lending dataset
2. (Choose 1 answer)

During cross-validation, which of the following dataset splits is used to tune the performance of a model?
A. Training set
B. Test set
C. Validation test
D. Unit test
3. (Choose 2 answers)

Which of the following are potential ethical benefits of conducting pilot testing prior to the release of emerging technologies? (Select two.)
A. It certifies that a product's developers are legally accountable.
B. It minimizes the impact of any ethical issues which may emerge.
C. It demonstrates that the system is functional as early as possible in order to satisfy senior management and investor stakeholders.
D. It reveals any gaps in accountability or auditability.
4. (Choose 1 answer)

Why is it important to consider Goodhart's law?
A. A model might optimize one metric at the expense of others.
B. A model might be ineffective in its application domain.
C. A model might change its task from classification to regression, or vice versa.
D. A model very famous and can be apply to all domain.
5. (Choose 2 answers)

Which of the following elements should be communicated with end users when negotiating ethical access to a user's data? (Select two.)
A. The monetary value of that data
B. An offer of money or other incentives to relinquish user's rights
C. The intention behind a certain activity or request
D. A request for consent to perform an action
6. (Choose 1 answer)

What are the limitations of training and testing an AI model in a specific context?
A. Models may perform poorly when applied to different contexts
B. Models require less computational power in limited contexts
C. Models reduce training data requirements when context is specified
D. Models automatically improve when tested in the same context
7. (Choose 1 answer)

What does the acronym RACI stand for?
A. Relatable, Agreeable, Commendable, Irascible
B. Responsible, Accessible, Consulted, Improvised
C. Redesignable, Accessible, Collected, Informative?
D. Responsible, Accountable, Consulted, Informed
8. (Choose 1 answer)

What are the essential steps involved in creating personas for user-centric marketing?
A. Define target audience, conduct research, build fictional personas based on insights, validate personas
B. Identify privacy risks, analyze data, implement policies, conduct training
C. Conduct surveys, run focus groups, implement findings, analyze results
D. Develop marketing strategies, execute campaigns, evaluate outcomes, adjust tactics
9. (Choose 1 answer)

Which of the following is a recommended security practice for machine learning datasets?
A. Ensure your team is full of different perspectives
B. Enact a sound data governance structure
C. Perform threat modeling with beneficial algorithms
D. Create a chain of command
10. (Choose 1 answer)

You're training a model to classify whether or not a bridge is likely to collapse given several factors. You have a dataset of thousands of existing bridges and their attributes, where each bridge is labeled as having collapsed or not collapsed. Only a handful of bridges in the dataset are labeled as having collapsed—the rest are labeled as not collapsed. Given your goal of minimizing bridge collapse and the severe harm it can cause, which of the following metrics would be most useful for evaluating the model?
A. Accuracy
B. Confusion matrix
C. Precision
D. Recall
11. (Choose 1 answer)

Which of the following describes extrajudicial judgment?
A. A ruling or scoring made outside of a court.
B. Transferring defendants from one jurisdiction to another for trial.
C. Applying law from one jurisdiction within another.
D. Unusual interpretation of law outside of its typical meaning.
12. (Choose 1 answer)

Differential privacy works by adding what to a dataset?
A. filters
B. fairness scoring
C. noise
D. k-anonymity
13. (Choose 1 answer)

When it comes to protecting private data, which of the following could be the root cause of a problem?
A. The data was stored insecurely.
B. The data was leaked online.
C. The data was deleted from storage.
D. The data was broken.
14. (Choose 1 answer)

What is the role of interdisciplinary collaboration in tech ethics?
A. It's unnecessary complexity
B. It brings diverse expertise to ethical challenges
C. It slows down development processes
D. It creates conflicting viewpoints only
15. (Choose 1 answer)

Which of the following describes the technique of homomorphic encryption?
A. A method of swapping public and private keys between hosts.
B. A method of cryptography that hides data inside other data.
C. A method of performing operations on encrypted data.
D. A method of mitigating biased algorithms through enhanced secrecy.
16. (Choose 1 answer)

What role does communication play in ethical leadership?
A. Preventing conflict resolution
B. Encouraging silence during ethical issues
C. Building trust and ethical alignment
D. Reducing transparency in strategy
17. (Choose 1 answer)

What can happen if users are excluded from AI impact discussions?
A. They contribute less
B. Systems may not reflect public values
C. Faster deployment
D. Lower training costs
18. (Choose 2 answers)

Which approach promotes inclusive communication?
A. Using a single communication channel
B. Considering diverse cultural and accessibility needs
C. Focusing only on majority perspectives
D. Standardizing all messages
19. (Choose 2 answers)

Which of the following are important qualities of a press release? (Select two.)
A. It's in a digital format
B. It's non-interactive
C. It's delivered on a single platform
D. It's concise
20. (Choose 1 answer)

What is a challenge of explainability in complex AI systems?
A. It makes models faster
B. Users don't care about explanations
C. Black-box algorithms are hard to interpret
D. It improves transparency instantly
21. (Choose 2 answers)

How can sentiment analysis be useful in social media communication?
A. It can shape the content and delivery of the message based on users' perceptions of the organization or technology.
B. It can classify customers as loyal or disloyal, helping you target customers differently in your communications.
C. It can determine which users have the largest influence on your reputation, and therefore which users you should communicate with.
D. It can determine how to construct a message that influences users to be more positive about the organization or technology.
22. (Choose 2 answers)

Which of the following are useful strategies for communicating ethical risks to the organization? (Select two.)
A. Communicate information through a single voice
B. Disseminate information in multiple formats
C. Communicate information in real time
D. Provide information only when solicited
23. (Choose 1 answer)

What is the purpose of a holding statement in crisis and media communications?
A. To order the business to temporarily halt certain operations
B. To retain the services of specific media outlets
C. To pre-write communications before they are needed
D. To prevent personnel from communicating prematurely
24. (Choose 2 answers)

Which of the following communication tactics help uphold transparency and explainability from the customer perspective? (Select two.)
A. Providing prompt responses to customer questions about the decisions made by data-driven systems.
B. Providing information about data-driven systems in simple, easy-to-understand terms.
C. Providing in-depth white papers on the algorithms that underlie data-driven systems.
D. Providing guidelines for how customers can keep their private data secure when using data-driven systems.
25. (Choose 1 answer)

What role does social media play in promoting products and building brand reputation?
A. Facilitates direct engagement with consumers
B. Guarantees higher sales numbers
C. Eliminates the need for customer feedback
D. Focuses solely on aesthetic appeal
26. (Choose 2 answers)

Why are deep learning and machine learning called narrow? Select two options.
A. Narrow can only perform the specific tasks it was designed to do.
B. Narrow AI can't handle situations that its training data didn't prepare it for.
C. Narrow AI is dependent on hand-coding of its algorithms.
27. (Choose 1 answer)

The optimistic view of general AI could be accurately summarized as AI as a ____
A. utility
B. peace-keeping tool
C. human right
D. weapon
28. (Choose 2 answers)

Which of the following are examples of how AI can limit human autonomy? (Select two.)
A. Weapon systems can limit human autonomy as humans may not have decision-making capability or understand the decision making.
B. AI systems might impact certain vulnerable groups such as the elderly and children differently than the rest of the population, which could limit those groups' autonomy.
C. AI systems can assist individuals with automated, repetitive, or dangerous tasks.
D. AI systems can perform tasks that humans cannot, such as processing millions of data records in a matter of seconds.
29. (Choose 1 answer)

For a model to clean, parse, and self-train it's own dataset while remaining impartial, the model needs
A. more powerful computing algorithms to auto-scrub data
B. a list of bias and domain tests to run and adjust for
C. 10x the amount of data
D. a test for recency bias
30. (Choose 1 answer)

Which of the following best distinguishes between responsibility, accountability, and liability?
A. Responsibility is the duty to complete a task, accountability is ownership of outcomes, and liability is legal obligation for any potential damages
B. Responsibility is managing outcomes, accountability involves delegating tasks, and liability is financial responsibility
C. Responsibility and accountability are synonymous, while liability is the duty to perform tasks
D. Liability and accountability are the same, whereas responsibility is irrelevant in business
31. (Choose 3 answers)

Which of the following, by itself, qualifies as personally identifiable information (PII)?
A. System events added to a log
B. A user's home address
C. Temperature readings for an office building
D. A user's customer ID in an online ordering system
32. (Choose 3 answers)

Which of the following are actions that can help combat implicit bias? (Select three.)
A. Obtaining your information from the same one or two media sources that your family and friends access.
B. Cultivating awareness of your own biases.
C. Interacting with diverse groups of people.
D. Exposing yourself to “counter-stereotypical” examples.
33. (Choose 1 answer)

Model inputs of address with "City + State" as separate inputs from a dataset would violate which accuracy guideline?
A. Domain expertise
B. No correlating data
C. First principles
D. Objective summarization
34. (Choose 1 answer)

Which type of bias causes people to trust an automated decision-making system (ADS) over a human's decision?
A. Implicit bias
B. Automation bias
C. Complacency bias
D. Confirmation bias
35. (Choose 1 answer)

A predictive model's false negative result can be defined as
A. the predicted result was positive, and the actual result was positive
B. the predicted result was negative, and the actual result was negative
C. the predicted result was positive, and the actual result was negative
D. the predicted result was negative, and the actual result was positive
36. (Choose 2 answers)

How can confirmation bias impact us socially? (Select two.)
A. It can lead to a diverse set of friends.
B. It can prevent us from being social with other people.
C. It can impede socio-political cooperation.
D. It can lead to groupthink, which can in turn halt forward progress.
37. (Choose 1 answer)

Which emerging technology area requires the most immediate ethical attention?
A. All emerging technologies require ethical consideration
B. Only artificial intelligence
C. Only biotechnology
D. Only robotics
38. (Choose 1 answer)

In a complex learning function, we will understand the ____ but not the ____
A. causal link, correlation
B. algorithm, cause and effect
C. inputs/outputs, algorithm
D. input data, output data
39. (Choose 1 answer)

What is the ultimate goal of ethics in emerging technology?
A. Slow down technological progress
B. Ensure technology serves humanity's best interests
C. Increase regulatory compliance costs
D. Make development more complex
40. (Choose 1 answer)

"Companies have an obligation to their shareholders" is part of a view that sees artificial intelligence as
A. an overall good for humanity, no matter the consequences
B. just another tool that accelerates research, like online advertising
C. a gimmick for enterprises, unless general intelligence is achieved
D. a harmful tool that will bring about the end of capitalism
41. (Choose 1 answer)

What is the impact of biased algorithms on decision-making?
A. They can lead to unfair and discriminatory outcomes
B. They enhance the precision and accuracy of decisions
C. They ensure all stakeholders are considered equally
D. They guarantee privacy and security in all data processes
42. (Choose 2 answers)

What is one of the recommendations for governments?
A. Promote human values.
B. Promote transparency in AI.
C. Promote international cooperation.
D. Promote collaboration.
43. (Choose 2 answers)

Which of the following ethical domains does the IEEE 7000 series explore? (Select two.)
A. Emulated empathy
B. Machine-readable privacy terms
C. Fair competition
D. Personnel safety
44. (Choose 1 answer)

A model that equalizes the number of mistakes it makes for each subgroup to reduce harm is deciding on
A. equality of true outcomes
B. equality of false negatives
C. equality of training data
D. equality of prediction bias
45. (Choose 1 answer)

Your business handles the personal data of California residents. Which of the following regulations would enable a resident to request that their data be deleted from your company's files?
A. PCI DSS
B. OECD Privacy Guidelines
C. COPPA
D. CCPA
46. (Choose 1 answer)

Which international standard provides guidelines for AI governance?
A. ISO 27001
B. IEEE Standards
C. ISO/IEC 42001
D. HIPAA
47. (Choose 1 answer)

If you are attempting to build a new framework for the research and development (R&D) of AI, which of the following frameworks might you look at first for its emphasis in this area?
A. The Beijing AI Principles
B. The American Medical Association's definition of artificial intelligence
C. The G20 AI Principles
D. The Montreal Declaration for a Responsible Development of Artificial Intelligence
48. (Choose 1 answer)

Why is the question of robot rights and emancipation one that isn't as important as addressing issues of bias, privacy, transparency, and other principles discussed in the various ethical frameworks?
A. These rights necessitate that robots become sentient entities, which is currently not feasible.
B. Humans are anthropocentric and don’t want to extend rights to other sentient entities.
C. There is no legal precedent for granting rights to entities that are not humans.
D. Robots are mechanical instruments and therefore don’t deserve to have rights.
49. (Choose 1 answer)

Which of the following software development principles is essential in the real-world deployment of AI-enabled software applications in critical scenarios like self-driving cars?
A. Robustness to adversarial examples
B. Architectural design analysis
C. Continuous integration and deployment of patch updates
D. Version control of the AI models deployed
50. (Choose 1 answer)

What does the GDPR primarily regulate?
A. Artificial intelligence development
B. Data protection and privacy
C. Software quality assurance
D. Network security protocols
51. (Choose 1 answer)

How can a leader promote ethical behavior across an organization?
A. By making decisions in isolation
B. By enforcing silence on ethical issues
C. By modeling ethical behavior and encouraging open dialogue
D. By limiting diversity in teams
52. (Choose 1 answer)

Which of the following might be a risk of not having an ethical organizational culture?
A. Employees might be less interested in their careers.
B. Employees might face cognitive dissonance.
C. Employees will experience a faster approval process to get a product launched.
D. Employees are more likely to express their opinions about the company.
53. (Choose 1 answer)

How can an organization promote an ethical culture from the top down?
A. Leadership actively models ethical behavior
B. Emphasizing profit margins over ethical considerations
C. Ignoring unethical behavior if it increases efficiency
D. Relying solely on external audit reports
54. (Choose 1 answer)

Which of the following is a problematic reason to pilot a code of ethics prior to widespread deployment?
A. The pilot phase helps to identify gaps in the current code of ethics.
B. The pilot phase helps to develop awareness and buy-in from the relevant stakeholders in the organization.
C. The pilot phase demonstrates the seriousness of the organization with respect to its ethical responsibilities.
D. The pilot phase helps to stress test a code of ethics in a real-world situation.
55. (Choose 1 answer)

Why is fair competition an important ethical consideration for data-driven technology businesses?
A. Fair competition prevents predatory pricing.
B. Fair competition builds user trust in technology as a whole.
C. Fair competition provides products choice to users.
D. Fair competition provides service choice to users.
56. (Choose 1 answer)

What is one outcome of building an ethical culture within an organization?
A. Reduced compliance with regulation
B. Higher trust among stakeholders
C. Slower development processes
D. More hierarchical power structure
57. (Choose 1 answer)

What makes ethical leadership crucial in the data-driven age?
A. It focuses solely on profit
B. It empowers ethical culture and trust
C. It allows ignoring stakeholders
D. It prevents communication
58. (Choose 1 answer)

Which expression best characterizes what a company or organizational culture is?
A. "This is how we want this done."
B. "This is how things are done around here."
C. "It is what it is."
D. "This is what we're told to do."
59. (Choose 1 answer)

What is the most effective way to measure ethical implementation success?
A. Financial metrics only
B. Combination of stakeholder feedback, incident tracking, and outcome assessment
C. Legal compliance scores only
D. Technical performance metrics only
60. (Choose 1 answer)

How can fair competition in data-driven technologies foster innovation and trust?
A. By encouraging diverse ideas and approaches
B. By limiting access to new technologies
C. By enforcing uniformity in data processing methods
D. By prioritizing profit over ethical standards