An opinion piece against AI hype. The writer calls it 'AI solutionism': governments assume a neural network will fix any problem. They ignore missing data, missing skills and security risks, and the cancer-treatment and court-sentencing failures show what happens next.
The first multiple-choice block (27-29) uses paragraphs A, B and D. The summary (30-35) pulls from E, F, K and L. YES / NO / NOT GIVEN (36-39) sits in G to I. Question 40 asks about the passage as a whole.
Multiple choice (writer's purpose)
Questions 27-29
How to attack this block: These ask what the writer is doing, not what the paragraph says. Paragraph A lists AI benefits, but only as what the headlines claim. That's why the answer to 27 is about a public perception, not about benefits.
Question 27
Answer: BParagraph A
What is the writer's purpose in the opening paragraph?
There seems to be no shortage of sensationalist news about how AI could cure diseases… Just looking at the media headlines, you might think
Question says → Passage says
a public perceptionhysteria about the future of artificial intelligence (AI) is everywhere / you might think
of AIsensationalist news / the media headlines
Why: Every claim in paragraph A is filtered through the media: 'hysteria', 'sensationalist news', 'media headlines', 'you might think'. The writer is describing how the public sees AI, not giving their own view, so B.
The trap: C is the strongest distractor because the paragraph does list cures, innovation and creativity. But they're what the sensational news claims, and the writer clearly doubts them. D fails for the same reason: 'infiltrated every aspect of society' is only what the headlines make you think.
Question 28
Answer: AParagraph B, last sentence
What does the writer do when discussing AI solutionism in paragraph B?
this mindset actually jeopardises the value of machine intelligence by disregarding important AI safety principles and setting unrealistic expectations
Question says → Passage says
points out a riskjeopardises / disregarding important AI safety principles
Why: 'Jeopardises' means puts in danger. The writer names the danger of solutionism: it ignores safety principles and sets unrealistic expectations. That's pointing out a risk, so A.
The trap: D is tempting because the paragraph opens by admitting AI has 'opened up a wealth of promising opportunities'. That's a concession about AI in general, not a weighing of solutionism's pros and cons. B and C belong to paragraph C, where Silicon Valley is mentioned.
Question 29
Answer: CParagraph D, last sentence
What does paragraph D suggest about many politicians?
they fail to realise the complexity around deploying advanced machine learning systems in the real world
Question says → Passage says
be unaware offail to realise
the challenges of implementingthe complexity around deploying
national AI initiativesnational AI initiatives / advanced machine learning systems
Why: 'Fail to realise' is 'be unaware of', and 'the complexity around deploying' systems is 'the challenges of implementing' them. C is a close paraphrase.
The trap: D picks up the 'technological race to dominate', but the writer reports that governments are in this race; it doesn't say politicians are too keen. A is the reverse of the writer's view: politicians overrate AI's potential, they don't underrate it.
Summary completion (word list A-I)
Questions 30-35
How to attack this block: Three decoys among nine options. 'Confidentiality' sits next to the right idea for 32, because data access and privacy feel related. What the passage says is 'special permissions', and that means approval.
Question 30
Answer: FParagraph E, last two sentences
Besides solutions, what do politicians think adding a neural network will promote?
adding a neural network to a system of government does not mean it will be instantaneously more inclusive or fair
Question says → Passage says
the mere addition of a neural networksimply adding a neural network / adding a neural network
promote equalitymore inclusive or fair
Why: A system that is more 'inclusive or fair' treats people more equally, so the idea is equality (F). The summary's 'mere addition' mirrors 'simply adding'.
The trap: A, 'reliability', fits 'promote' grammatically, but the passage is about fairness, not about whether the system works dependably.
Question 31
Answer: GParagraph F, first sentence
What haven't most public sector organisations set up to handle the data?
the public sector typically does not have the appropriate data infrastructure to support advanced machine learning
Question says → Passage says
have not set up the necessarydoes not have the appropriate
frameworkdata infrastructure
Why: Infrastructure is the underlying structure that supports a system, which is what a framework is. G.
The trap: E, 'computers', is a concrete guess at what infrastructure means, but the passage never mentions hardware. 'Funding' (B) is also never mentioned.
Question 32
Answer: IParagraph F, fourth sentence
What does each person need before accessing data spread across departments?
data is spread across different government departments that each require special permissions to be accessed
Question says → Passage says
needs ___ to accessrequire special permissions to be accessed
spread across different departmentsspread across different government departments
Why: A permission is official approval to do something. The summary keeps 'spread across different departments' almost word for word, so you know you're in the right place. I.
The trap: H, 'confidentiality', is the decoy. Restricted data feels private, so confidentiality seems related, but the passage talks about permission to access, not keeping information secret.
Question 33
Answer: CParagraph F, last sentence
What do few public sector workers have that they'd need to use machine intelligence fully?
Above all, the public sector typically lacks the human talent with the right technological capabilities to fully reap the benefits of machine intelligence.
Question says → Passage says
the main problemAbove all
few public sector employees havethe public sector typically lacks the human talent
skillsthe right technological capabilities
take full advantage offully reap the benefits of
Why: 'Technological capabilities' in people means skills. 'Above all' signals the main problem, just as the summary says. C.
The trap: G, 'framework', has just been used for the infrastructure gap. This sentence is about people, so it needs a human quality.
Question 34
Answer: AParagraph K, third sentence
What did the medical experts have little faith in?
Even though it was developed to deliver the best recommendations, human experts found it hard to trust the machine.
Question says → Passage says
had little faith infound it hard to trust
its ___the machine / the best recommendations
Why: If you can't trust a machine, you doubt that it will give dependable results, and that is its reliability. A.
The trap: D, 'prejudices', belongs to the courts example in paragraph L. Nothing in K suggests the cancer programme was biased.
Question 35
Answer: DParagraph L, last sentence
What did the US court algorithms reflect and magnify?
The system was found to amplify structural racial discrimination and was later abandoned.
Question says → Passage says
magnifiedamplify
the existing prejudicesstructural racial discrimination
Why: 'Amplify' is 'magnify', and racial discrimination is a form of prejudice. 'Structural' tells you it was already built into the system, which is the summary's 'existing'. D.
The trap: F, 'equality', is the opposite idea, and A, 'reliability', is the medical example. Only 'prejudices' describes discrimination.
YES / NO / NOT GIVEN (writer's claims)
Questions 36-39
How to attack this block: The writer quotes two critics approvingly and complains that security is overlooked. Once you see whose side the writer is on, the NO answers fall out: the writer praises Russell, and says security is often ignored.
Question 36
Answer: NOParagraph G, second sentence
Does the writer think Stuart Russell's ideas about AI are unworkable?
has long advocated a more sensible and realistic approach that focuses on simple everyday applications of AI
Question says → Passage says
Stuart Russell's proposalshas long advocated a more sensible and realistic approach
impracticalsensible and realistic / simple everyday applications
Why: The writer describes Russell's approach as 'sensible and realistic', and focused on 'simple everyday applications'. Impractical is the opposite, so NO.
The trap: NOT GIVEN tempts people because the writer is reporting Russell's view. But calling it 'sensible and realistic' is the writer's own description, and 'Similarly' in the next sentence shows the writer agrees with him.
Question 37
Answer: NOT GIVENParagraph G, first and last sentences
Has Rodney Brooks been unfairly criticised by AI supporters?
the sensationalism over AI has attracted many critics… Similarly, Rodney Brooks, professor of robotics at Massachusetts Institute of Technology, writes
Question says → Passage says
attracted criticismhas attracted many critics
Why: Brooks is one of the critics of the AI hype. The passage reports what he writes, but nothing about how AI supporters responded to him, fairly or unfairly. NOT GIVEN.
The trap: 'Attracted many critics' is in the paragraph, so it feels like the statement's 'attracted criticism'. But it's the AI sensationalism that attracted critics, and Brooks is one of those critics, not their target.
Question 38
Answer: NOParagraph H, last sentence
Is protecting AI systems now always considered when they're installed?
Still, AI security remains an often overlooked topic when machine learning systems are installed.
Question says → Passage says
the need to protect AI systemsAI security
is always taken into accountremains an often overlooked topic
when they are set upwhen machine learning systems are installed
Why: 'Often overlooked' means often ignored, which directly contradicts 'always taken into account'. NO.
The trap: The sentence before says researchers have warned about security, which might make you think the warnings were heeded. 'Still' tells you they weren't.
Question 39
Answer: YESParagraph I
To get AI's benefits and limit its harm, do we need to look at people's worries about it?
If we are to reap the benefits and minimise the potential harms of AI… This means we need to have a discussion about AI ethics and the distrust that many people have towards machine learning.
Question says → Passage says
in order to benefit from AI and minimise the harmsIf we are to reap the benefits and minimise the potential harms of AI
explore people's concernshave a discussion about… the distrust that many people have
Why: Having a discussion about people's distrust is exploring their concerns, and the writer says 'we need to' do it. YES.
Multiple choice (best subtitle)
Question 40
How to attack this block: A subtitle question tests the whole argument. The last paragraph says there is no AI solution for everything. The right option says the same thing. The distractors either go too far (don't invest at all) or say the opposite.
Question 40
Answer: BParagraph M (and the passage as a whole)
Which subtitle sums up the whole passage?
These examples demonstrate that there is no AI solution for everything.
Question says → Passage says
may not be the answer to our problemsthere is no AI solution for everything / not every issue is best addressed by applying machine intelligence
Why: The whole article argues against 'AI solutionism', the belief that AI can solve everything, and the conclusion says it outright. B matches that argument.
The trap: C goes too far: the writer criticises unrealistic expectations, not investment itself, and paragraph M speaks to people boosting investment without telling them to stop. A sounds positive and practical, but the article is mainly a warning, not a how-to guide.
Option C is tempting because paragraph A lists cures and creativity. They're framed as sensational headlines, so the writer is describing how the public sees AI (B).
Each department needs special permissions before the data can be opened. That's approval (I). 'Confidentiality' is the decoy.
Words from this passage worth keeping
infiltrate(A)
gradually get into every part of something
jeopardise(B)
put something at risk of being harmed or lost
utopian(C)
imagining a perfect world, often unrealistically
burgeoning(D)
growing quickly
reap the benefits(F)
get the advantages that come from something
advocate(G)
publicly support an idea or approach
susceptible to(H)
easily affected or harmed by something
amplify(L)
make something stronger or bigger
Sit this test on Langujet, then come back
The full Cambridge 21 Test 2 reading paper: all three passages, 60 minutes, marked the second you finish with a review of every answer. Already done it? Take it again and see if the same traps still get you.
Cambridge IELTS 21, Test 2, Reading Passage 3. The passage and questions belong to Cambridge University Press and are not reproduced here. We quote short extracts only as evidence for our own commentary, and the answers are published for self-marking. Get the book for the full text, or sit the test on Langujet.
FAQ
Artificial Intelligence: quick answers
"Artificial Intelligence" is Reading Passage 3 of Cambridge IELTS 21, Test 2. It covers questions 27-40 (14 marks).
Questions 27-29: Multiple choice, questions 30-35: Summary completion, questions 36-39: YES / NO / NOT GIVEN and question 40: Multiple choice.
1 YES, 2 NO and 1 NOT GIVEN.
Our review highlights questions 36, 27 and 32. Each one is explained in "Questions worth a second look" above.
