Đề thi thử VSTEP Listening số 19
Thời gian: 40 phút | Tổng số câu: 35 câu trắc nghiệm
Hướng dẫn: Chọn đáp án đúng nhất (A, B, C hoặc D) cho mỗi câu hỏi. Bạn chỉ được nghe MỘT lần.
PART 1 — Short Announcements (Câu 1–8)
Nghe 8 đoạn thông báo ngắn. Mỗi đoạn có 1 câu hỏi.
Câu 1. What is the company announcing regarding AI?
A. A complete ban on AI in the workplace
B. A new AI ethics policy requiring human oversight for all automated decisions affecting customers
C. A new product launch
D. Office closure for renovation
Câu 2. What new AI system is being deployed?
A. A robot nurse
B. An AI-assisted diagnostic tool that helps radiologists detect tumors earlier and more accurately
C. A hospital billing system
D. A new cafeteria menu
Câu 3. What is the university requiring students to do?
A. Avoid using all technology
B. Complete a new course on AI literacy, ethics, and responsible use before graduation
C. Submit handwritten essays only
D. Pay additional fees
Câu 4. What new service is the firm offering?
A. Manual document review only
B. AI-assisted contract review and legal research to reduce costs and time for clients
C. A new office location
D. Free legal advice
Câu 5. What regulation is being introduced?
A. A ban on all AI development
B. New transparency requirements for AI systems used in hiring, lending, and other consequential decisions
C. Free AI for all citizens
D. Mandatory AI use in schools
Câu 6. What change is being made?
A. Removing all online services
B. Using AI to accelerate claims processing, with human review for complex cases
C. Increasing premiums by 200 percent
D. Closing branches
Câu 7. What opportunity is being offered?
A. Free flights
B. A fellowship program for researchers studying AI safety and alignment
C. Discounted hotels
D. A new sports program
Câu 8. What is the studio announcing?
A. A complete ban on AI in filmmaking
B. A new policy allowing AI-assisted special effects while protecting human writers and actors' rights
C. A new streaming service
D. A new restaurant
PART 2 — Conversations (Câu 9–20)
Nghe 3 đoạn hội thoại. Mỗi đoạn có 4 câu hỏi.
Câu 9. What task does the woman want to use AI for?
A. Replacing all employees
B. Automating routine data analysis to free up time for more creative strategic work
C. Writing personal emails
D. Ordering lunch
Câu 10. What concern does her colleague raise?
A. AI is too slow
B. The AI tools sometimes produce biased or inaccurate results, which could affect business decisions
C. AI is too expensive
D. AI is illegal
Câu 11. What does the woman suggest to address the concern?
A. Stop using AI entirely
B. Always have a human review AI outputs before making important decisions, and document any errors
C. Trust AI completely
D. Hire more people to do manual work
Câu 12. What do they agree to do?
A. Ban AI
B. Pilot the AI tool in one department, with clear guidelines and human oversight, then expand based on results
C. Tell no one about the experiment
D. Take a long vacation
Câu 13. What is the parent concerned about?
A. The school is too strict
B. Their child is using AI to complete assignments without learning the material
C. The school has too many activities
D. The food in the cafeteria
Câu 14. How does the teacher respond?
A. Bans all technology in the classroom
B. Acknowledges the concern but explains that AI literacy is now an essential skill, and teaches students how to use AI as a learning tool ethically
C. Refuses to discuss the issue
D. Suggests the child switch schools
Câu 15. What does the teacher suggest the parent do at home?
A. Ban all computers
B. Have open conversations about the difference between using AI to learn versus using AI to cheat, and set clear expectations
C. Pay for expensive tutors
D. Ignore the issue
Câu 16. How does the parent feel after the conversation?
A. Angry and upset
B. More understanding and willing to work with the teacher to set appropriate guidelines
C. Indifferent
D. Determined to remove the child from school
Câu 17. What is the main concern about advanced AI systems?
A. They are too slow
B. They may develop goals misaligned with human values and become difficult to control
C. They are too expensive
D. They are illegal
Câu 18. What is "alignment" in the context of AI safety?
A. A type of machine learning
B. The challenge of ensuring AI systems act in accordance with human intentions and values
C. A type of hardware
D. A government agency
Câu 19. What approach do they discuss to improve AI safety?
A. Banning all AI
B. Developing techniques for AI transparency, interpretability, and robust human oversight mechanisms
C. Letting AI make all decisions
D. Reducing AI capabilities
Câu 20. What is their overall view on AI safety research?
A. It is a waste of time
B. It is critically important and underfunded, deserving more attention and resources
C. It is too late to act
D. It is only relevant for military AI
PART 3 — Lectures (Câu 21–35)
Nghe 3 bài giảng. Mỗi bài có 5 câu hỏi.
Câu 21. What is one of the most pressing ethical concerns about AI, according to the speaker?
A. AI is too expensive
B. Algorithmic bias, where AI systems perpetuate or even amplify existing societal biases in hiring, lending, criminal justice, and other areas
C. AI is too slow
D. AI is not used enough
Câu 22. Why does bias occur in AI systems, according to the speaker?
A. AI is designed to be biased
B. Bias reflects the biases present in the training data, often collected from historical sources that contain societal prejudices
C. Bias is added intentionally by developers
D. Bias only happens in image recognition
Câu 23. What example does the speaker give of algorithmic bias?
A. A music recommendation
B. A hiring tool that favored male candidates because it was trained on resumes from a male-dominated industry
C. A weather prediction model
D. A cooking app
Câu 24. What does the speaker suggest as a solution?
A. Banning AI
B. Carefully curating training data, auditing AI systems for bias, ensuring diverse teams develop AI, and ongoing monitoring
C. Using only one type of AI
D. Ignoring bias
Câu 25. Why is transparency important in AI, according to the speaker?
A. It is not important
B. It allows users and regulators to understand how AI systems make decisions, identify potential issues, and hold developers accountable
C. It slows down AI
D. It is only for researchers
Câu 26. What is generative AI, according to the speaker?
A. A type of database
B. A category of AI systems that can create new content—including text, images, audio, and code—based on patterns learned from training data
C. A type of computer hardware
D. A type of internet connection
Câu 27. What breakthrough made modern generative AI possible, according to the speaker?
A. The invention of the internet
B. The transformer architecture, introduced in 2017, which allowed AI models to be trained on massive amounts of data and to generate coherent, contextually relevant outputs
C. A new type of battery
D. Faster internet speeds
Câu 28. How are generative AI tools like ChatGPT typically trained, according to the speaker?
A. On small datasets
B. On enormous amounts of text from books, articles, websites, and other sources, learning patterns and relationships in language
C. Only on academic papers
D. On user data only
Câu 29. What is one concern about generative AI, according to the speaker?
A. It is too fast
B. It can produce convincing but false information, raise questions about copyright and intellectual property, and disrupt many existing jobs
C. It is too accurate
D. It is only available in English
Câu 30. What does the speaker identify as a key challenge?
A. Making AI more expensive
B. Balancing the tremendous creative and productive potential of generative AI with concerns about misuse, accuracy, and societal impact
C. Eliminating AI
D. Banning AI development
Câu 31. How is AI being used in healthcare, according to the speaker?
A. Replacing all doctors
B. Assisting in diagnosis, drug discovery, personalized treatment plans, and predicting patient outcomes
C. Only for billing
D. For hospital administration only
Câu 32. What is one specific example of AI in healthcare that the speaker mentions?
A. AI that can predict disease outbreaks
B. AI systems that can detect diabetic retinopathy from eye scans with accuracy comparable to human specialists
C. AI for scheduling appointments
D. AI for hospital food services
Câu 33. What concern does the speaker raise about AI in healthcare?
A. AI is too accurate
B. Questions about accountability when AI systems make errors, and concerns about patient data privacy and security
C. AI is too fast
D. AI is banned
Câu 34. According to the speaker, why won't AI replace doctors in the near future?
A. AI is illegal
B. Medicine requires empathy, complex ethical judgment, and the human connection between doctor and patient, which AI cannot replicate
C. AI is too expensive
D. Doctors are protected by law
Câu 35. What does the speaker say is the most likely future of AI in healthcare?
A. AI will replace all healthcare workers
B. AI will augment and assist healthcare professionals, making them more effective and enabling better patient care
C. AI will be banned in healthcare
D. AI will be used only for research