IELTS Academic Reading Practice Test 9

A. Artificial intelligence is transforming healthcare at a pace that would have seemed inconceivable just a decade ago. Machine learning algorithms can now analyse medical images with accuracy matching or exceeding that of experienced radiologists, predict patient deterioration hours before clinical signs become apparent, and identify potential drug candidates in a fraction of the time required by traditional pharmaceutical research. The global AI in healthcare market, valued at approximately $15 billion in 2023, is projected to exceed $180 billion by 2030.

B. Diagnostic imaging represents the most mature application of AI in clinical medicine. Deep learning algorithms trained on millions of labelled images have demonstrated remarkable performance in detecting conditions including diabetic retinopathy, skin cancer, breast cancer on mammography, and lung nodules on chest X-rays. In 2020, Google Health published a study in Nature showing that their AI system detected breast cancer on mammograms with greater accuracy than individual radiologists, reducing both false positives and false negatives. However, the system was trained primarily on data from the United States and United Kingdom, raising questions about its performance in populations with different demographic characteristics.

C. Drug discovery is another area where AI is making significant inroads. The traditional process of developing a new pharmaceutical compound from initial discovery to market approval takes an average of 12 to 15 years and costs approximately $2.6 billion. AI-driven approaches can dramatically accelerate the early stages of this process by predicting how molecular structures will interact with biological targets, identifying promising candidates from vast chemical libraries, and optimising lead compounds for desired properties. In 2023, the biotechnology company Insilico Medicine became one of the first to advance an AI-discovered drug candidate to Phase II clinical trials for idiopathic pulmonary fibrosis.

D. Predictive analytics powered by AI enables healthcare systems to identify patients at high risk of adverse outcomes before they occur. Electronic health record data, combined with real-time monitoring from wearable devices and bedside sensors, can be analysed by machine learning models to generate early warning scores for conditions such as sepsis, cardiac arrest, and hospital readmission. The Epic Sepsis Model, deployed across hundreds of US hospitals, uses patient data to predict the onset of sepsis up to six hours before clinical recognition, potentially enabling earlier intervention and improved survival rates.

E. Despite these promising developments, the deployment of AI in healthcare faces substantial challenges. Algorithmic bias, reflecting the biases present in training data, can lead to systematic disparities in care. A widely publicised 2019 study in Science revealed that a commercial algorithm used by US health systems to allocate healthcare resources was systematically underestimating the needs of Black patients, because it used healthcare spending as a proxy for health needs, a metric that reflected existing disparities in access to care rather than actual medical requirements.

F. Regulatory frameworks for medical AI remain in development. The US Food and Drug Administration has approved several hundred AI-enabled medical devices, but the regulatory pathway for continuously learning algorithms that update their predictions as new data becomes available is still being defined. Questions of liability when AI systems contribute to clinical errors, the interpretability of complex neural networks that function as "black boxes," and the protection of patient data used to train AI models all require careful resolution. The integration of AI into clinical workflows also demands significant investment in infrastructure, training, and change management within healthcare organisations that are often resistant to technological disruption.

Each year, an estimated 50 billion birds undertake migratory journeys, some spanning thousands of kilometres across continents and oceans. The Arctic tern holds the record for the longest migration of any animal, travelling approximately 71,000 kilometres annually between its Arctic breeding grounds and Antarctic wintering areas, effectively experiencing two summers per year. Bar-tailed godwits complete the longest non-stop flight, covering up to 12,000 kilometres from Alaska to New Zealand in approximately nine days without resting, eating, or drinking.

Birds navigate using a remarkable array of sensory mechanisms. The sun compass, which requires the bird to compensate for the sun's movement across the sky using an internal clock, provides directional information during daylight hours. At night, many species orient by the stars, using the rotation of the night sky around the celestial pole to determine north. Perhaps most remarkably, birds possess a magnetic sense that allows them to detect the Earth's magnetic field. Research has identified two potential mechanisms: magnetite crystals in the upper beak that function like a biological compass needle, and a quantum mechanical process in the eye involving cryptochrome proteins that may allow birds to literally see magnetic field lines.

Climate change is disrupting migratory patterns in multiple ways. Rising temperatures are causing many species to shift their breeding ranges northward, with European birds advancing their spring arrival dates by an average of two to three days per decade since the 1970s. However, the timing of insect emergence, upon which many migratory birds depend for feeding their young, is advancing more rapidly than bird arrival dates, creating a phenological mismatch that can reduce breeding success. In some cases, traditionally migratory populations are becoming partially resident, with individuals in mild winters choosing not to migrate at all.

Habitat loss along migratory routes, known as flyways, represents the most significant threat to migratory bird populations. The East Asian-Australasian Flyway, used by over 50 million waterbirds, has lost more than 65 percent of its intertidal mudflat habitat to coastal development, particularly in China and South Korea. The reclamation of the Saemangeum estuary in South Korea, completed in 2006, destroyed approximately 400 square kilometres of tidal flat that had supported hundreds of thousands of migratory shorebirds, contributing to dramatic population declines in species such as the spoon-billed sandpiper.

Conservation of migratory species requires international cooperation, as birds crossing national boundaries cannot be protected by any single country's legislation alone. The Convention on Migratory Species, also known as the Bonn Convention, provides a legal framework for coordinating conservation across range states. Bilateral agreements, such as the Migratory Bird Treaty between the United States, Canada, Mexico, Japan, and Russia, establish protections for shared species. Satellite tracking technology has revolutionised our understanding of migration routes, enabling researchers to identify critical stopover sites that require protection and to quantify the threats birds face at each stage of their journey.

Classical computers process information using bits that exist in one of two states: 0 or 1. Quantum computers exploit the principles of quantum mechanics to process information using quantum bits, or qubits, which can exist in a superposition of both states simultaneously. This fundamental difference enables quantum computers to perform certain types of calculations exponentially faster than any classical machine, regardless of its processing power. While a classical computer with n bits can represent only one of 2^n possible states at any given time, a quantum computer with n qubits can represent all 2^n states simultaneously, enabling massively parallel computation.

The theoretical foundations of quantum computing were laid in the 1980s by physicists Richard Feynman, who proposed that quantum systems could be efficiently simulated only by quantum machines, and David Deutsch, who described the first universal quantum computer. Peter Shor's 1994 algorithm for factoring large numbers demonstrated the potential of quantum computing to break widely used encryption systems, while Lov Grover's 1996 search algorithm showed that quantum computers could search unsorted databases quadratically faster than classical alternatives.

Building practical quantum computers presents extraordinary engineering challenges. Qubits are extremely fragile; they lose their quantum properties through a process called decoherence when they interact with their environment. Current approaches to maintaining coherence include cooling superconducting circuits to temperatures near absolute zero (approximately -273 degrees Celsius), trapping individual ions with electromagnetic fields, and encoding qubits in photons of light. Google's Sycamore processor, which in 2019 achieved what the company termed "quantum supremacy" by performing a specific calculation in 200 seconds that would take a classical supercomputer approximately 10,000 years, used 53 superconducting qubits cooled to 15 millikelvin.

Error correction remains the central technical barrier to useful quantum computation. Environmental noise causes errors in qubit operations at rates far exceeding those in classical transistors. Quantum error correction schemes require encoding each logical qubit across multiple physical qubits, with estimates suggesting that thousands of physical qubits may be needed for each error-corrected logical qubit. Current quantum processors with fewer than 1,500 physical qubits are therefore limited to relatively simple calculations, a stage of development often referred to as the Noisy Intermediate-Scale Quantum (NISQ) era.

The most promising near-term applications lie in molecular simulation, optimisation problems, and machine learning. Pharmaceutical companies are investing in quantum computing to simulate molecular interactions at a level of accuracy impossible for classical computers, potentially accelerating drug discovery. Financial institutions are exploring quantum algorithms for portfolio optimisation and risk analysis. Major technology companies including IBM, Google, Microsoft, and several well-funded startups are competing to build increasingly powerful quantum processors, with IBM targeting a 100,000-qubit system by 2033.

The implications for cryptography are particularly consequential. Most current internet security relies on the difficulty of factoring large prime numbers, a problem that Shor's algorithm could solve efficiently on a sufficiently powerful quantum computer. Recognising this threat, the US National Institute of Standards and Technology has been developing post-quantum cryptographic standards since 2016, with the first algorithms selected in 2022. The race to develop quantum-resistant encryption before large-scale quantum computers become available represents one of the most significant cybersecurity challenges of the coming decades.