Công nghệ & AI · Reading: Yes / No / Not Given (dạng khó)

SHOULD WE FEAR ARTIFICIAL INTELLIGENCE? Few technologies have been discussed as much as artificial intelligence. Some commentators warn that intelligent machines will soon take over most jobs and even threaten human survival. Others promise that AI will cure diseases and end poverty within a generation. In my view, both groups exaggerate. There is no doubt that machine learning has produced genuine breakthroughs. Systems that analyse medical scans are, in some studies, as accurate as experienced doctors, and translation software has improved at an unprecedented speed. However, these systems are narrow: a program that identifies skin cancer cannot drive a car or write a poem. The idea that one machine will soon match human intelligence in every area is, I believe, not supported by the evidence. That does not mean we should relax. The more immediate danger is not a super-intelligent robot but ordinary algorithms that make decisions about people's lives without anyone checking them. Algorithms already help decide who gets a loan, who is invited to a job interview and, in some countries, who receives police attention. If the data they learn from reflects past discrimination, they are likely to repeat it — and because their decisions look scientific, they may be harder to challenge than a biased human. The answer, I would argue, is not to ban AI but to regulate it sensibly. Companies should be required to explain how important decisions are made, and a human should always be able to review them. Such rules may slow innovation slightly, but that is a small price for public trust.

1. 1. The writer agrees that AI will take over most jobs soon.

   YNNG

2. 2. The writer accepts that machine learning has achieved real progress.

   YNNG

3. 3. The writer believes medical AI is more accurate than all doctors.

   YNNG

4. 4. The writer thinks translation software will soon replace human translators.

   YNNG

5. 5. According to the writer, ordinary algorithms are a more urgent risk than super-intelligent machines.

   YNNG

6. 6. The writer suggests that biased algorithms may be harder to challenge than biased people.

   YNNG

7. 7. The writer believes AI should be banned in the police.

   YNNG

8. 8. The writer thinks new rules would greatly reduce companies' profits.

   YNNG

1. 1. AI ___ to change most office jobs in the next ten years.

   A. likelyB. is likelyC. will likely toD. likely is

2. 2. Which sentence is the most certain?

   A. AI may replace some drivers.B. AI is likely to replace some drivers.C. AI will replace some drivers.D. AI might replace some drivers.

3. 3. Which sentence is the least certain?

   A. Robots are unlikely to replace nurses.B. Robots will not replace nurses.C. It is certain that robots won't replace nurses.D. Robots cannot replace nurses.

4. 4. Passage: 'AI may help teachers mark essays.' Statement: 'The writer believes AI will replace teachers.' →

   A. YesB. NoC. Not Given

5. 5. ___, privacy is the biggest risk of facial recognition.

   A. LikelyB. ArguablyC. AccuratelyD. Unlikely

1. 1. AI will probably change education. (likely) → AI
2. 2. It is possible that autonomous cars will reduce accidents. (may) → Autonomous cars
3. 3. Robots probably won't replace nurses soon. (unlikely) → Robots
4. 4. The algorithm seems to prefer men. (appears) → The algorithm

1. algorithm— chọn —an important new discoverybeing open about how decisions are madenever having happened beforeunfairly favouring one group over anotherpeople checking what a system doesa set of rules a computer follows to solve a problem
2. bias— chọn —an important new discoverybeing open about how decisions are madenever having happened beforeunfairly favouring one group over anotherpeople checking what a system doesa set of rules a computer follows to solve a problem
3. transparency— chọn —an important new discoverybeing open about how decisions are madenever having happened beforeunfairly favouring one group over anotherpeople checking what a system doesa set of rules a computer follows to solve a problem
4. breakthrough— chọn —an important new discoverybeing open about how decisions are madenever having happened beforeunfairly favouring one group over anotherpeople checking what a system doesa set of rules a computer follows to solve a problem
5. unprecedented— chọn —an important new discoverybeing open about how decisions are madenever having happened beforeunfairly favouring one group over anotherpeople checking what a system doesa set of rules a computer follows to solve a problem
6. human oversight— chọn —an important new discoverybeing open about how decisions are madenever having happened beforeunfairly favouring one group over anotherpeople checking what a system doesa set of rules a computer follows to solve a problem

1. 1. Facial recognition raises serious questions.(ethics)
2. 2. The of medical AI has improved.(accurate)
3. 3. Some hiring algorithms are against women.(bias)
4. 4. AI has the to transform healthcare.(potentially)
5. 5. vehicles may reduce accidents.(autonomy)

1. 1. What does the council claim?

   A. Crime will fall slightly.B. Crime will be cut by half.C. Crime will not change.D. Cameras are cheap.

2. 2. What does the woman think about the council's claim?

   A. It is certain.B. It is unlikely.C. It is proven.D. She has no opinion.

3. 3. What is the woman's main worry?

   A. costB. privacyC. accuracyD. slow trains

4. 4. Under what condition might she accept the cameras?

   A. if they are cheapB. with strict rules and human oversightC. if crime risesD. never

1. 1. Chatbots can generate computer code.

   TF

2. 2. Research shows chatbots always harm learning.

   TF

3. 3. Students who used chatbots to check ideas wrote much better essays.

   TF

4. 4. Researchers recommend banning chatbots in schools.

   TF