IELTS Academic Reading Practice Test 159

Few pieces of technology are as completely taken for granted as the small plastic disc screwed to a ceiling. In countries whose building regulations demand them, most homes now contain at least one smoke alarm, and their contribution to safety has been considerable. Fire services in several countries have recorded substantial falls in deaths from house fires since alarms became widespread in the 1970s, although the decline also reflects safer furniture, fewer smokers and stricter regulations. The alarm itself, however, is far older than its modern ubiquity suggests.

The first automatic fire detectors responded not to smoke but to heat. From the 1870s, automatic sprinkler systems used a fusible link, a small joint of solder designed to melt at a chosen temperature and so release water onto the fire below. In 1890 Francis Robbins Upton, a physicist who had worked with Thomas Edison, patented one of the first electric fire alarms, in which heat completed a circuit that operated a bell. Such devices worked well enough in factories and theatres, but they were slow to respond. A smouldering fire might burn for a long time before the air around the sensor grew hot enough to trigger the mechanism.

The decisive step towards detecting smoke itself came from an unexpected direction: research on gases rather than fires. By the 1930s, physicists knew that ionising radiation could be used to measure what was present in air, because the ions it creates carry a tiny electric current which other substances disturb. An often-repeated account describes a researcher building an instrument to warn of poisonous gas who left a cigarette smouldering beside it and found that the apparatus registered the smoke. Whatever the truth of that story, the principle was sound: particles of smoke cling to ions and reduce the current they carry.

Almost every ionization alarm made today contains a minute quantity of americium-241, an isotope with a half-life of 432 years. Sealed inside a small chamber, it emits alpha particles, fast-moving helium nuclei that strip electrons from air molecules as they travel. The result is a steady population of positive and negative ions, which a modest voltage draws towards an electrode, producing a current of a few billionths of an amp. When smoke enters, its particles attach to the ions and slow them, so the current falls; once it drops below a set threshold, the alarm sounds.

The alternative design works with light. In an optical, or photoelectric, alarm, a light-emitting diode projects a beam across a chamber whose walls are shaped so that no light reaches a photodiode placed at right angles to it. Air alone scatters almost nothing. Smoke is another matter: its particles deflect light in every direction, and some of it lands on the photodiode, which converts it into a signal. Optical devices of this general kind were demonstrated in the 1940s, but for two decades both designs remained costly equipment for industry.

What turned the alarm into a household object was the transistor. Early detectors used valves and mains wiring, which made them bulky and expensive to install. Transistorised circuits needed very little power, so a detector could run for a year on a small battery. The American engineer Duane Pearsall, whose company produced one of the first battery-powered domestic alarms in the mid-1960s, is often credited with the breakthrough, although prices kept alarms out of poorer households for years afterwards.

During the 1970s sales grew at an extraordinary rate. In the United States, fewer than one home in ten had an alarm at the start of the decade; by 1980 the figure was around three-quarters. Price was part of the explanation, since costs fell as production expanded, and publicity was another: fire services ran campaigns showing that most fatal fires happened at night, when occupants were asleep. Legislation added momentum, as several American states and a number of European countries began to require alarms in new dwellings, a rule that Britain introduced in the early 1990s.

Two problems have persisted. The first is the false alarm. Cooking fumes, steam from a shower and burnt toast all produce particles that resemble smoke closely enough, and an alarm that shrieks whenever a saucepan is left on the hob is soon silenced. Householders disable them by removing the battery or covering the device, and investigations of fatal fires repeatedly find that no working alarm was present.

Hush buttons and better shielding of the chamber have helped, but the tension between sensitivity and nuisance cannot be removed. The second problem concerns the choice of sensor: ionization alarms react more quickly to fast, flaming fires, while optical alarms respond sooner to the smouldering fires that kill most often at night, and neither is best in all circumstances.

Contemporary systems differ less in their sensors than in what they do with the signal. Mains-powered alarms can be interconnected by wire or radio, so that a fire in a basement sets off every unit in the house, including one outside a bedroom. Some models emit a spoken message or a low-frequency tone, which is easier for people with impaired hearing to detect. Disposal raises a modest environmental question, since the americium in an ionization alarm decays slowly, and some authorities ask householders to return used units to the supplier.

Shortly after dawn one winter morning, a box-shaped container ship slipped away from a quay at Porsgrunn in southern Norway and made the short crossing of a fjord to a fertiliser plant. The vessel, the Yara Birkeland, runs on batteries and carries bags of fertiliser; on many voyages it also carries nobody. A control centre on shore follows its progress through cameras and radar, ready to intervene if something goes wrong. The ship belongs to a class that regulators call maritime autonomous surface ships, or MASS: vessels able to operate with varying degrees of independence from the people who own and manage them.

The interest in such ships is easy to explain. Crews are expensive, since wages, cabins, food and the space they occupy all cost money, and on a short-sea route where a vessel spends as much time in port as at sea the saving can be substantial. Safety is a stronger argument. Industry estimates attribute the great majority of accidents at sea to human error, and fatigue among crews working long watches is a well-documented hazard. An automated vessel, its designers argue, does not grow tired and does not lose concentration.

Replacing human senses is harder than it sounds. A modern autonomous ship carries radar, a receiver for the Automatic Identification System that broadcasts the position of other vessels, daylight and thermal cameras, and increasingly lidar, which measures distance with pulses of laser light. Each sensor has weaknesses. Radar struggles to pick out small wooden boats in choppy water, cameras are defeated by fog and glare, and the identification system depends on other vessels transmitting accurate data, which not all of them do. The software must fuse these imperfect pictures into one model of the surroundings. It must also satisfy a legal requirement that has existed for well over a century: the obligation to keep a proper lookout.

Judgement is the next difficulty. The international regulations for preventing collisions at sea, known as COLREGs, were drafted for officers who could see one another's ships and read intent from a masthead light or a change of course. Their language is deliberately flexible: a vessel must take substantial action to keep clear, and must do so in ample time. Writing software to apply such phrases is not simply a coding problem, because the rules assume that a competent human will decide what they mean in a particular encounter. A machine has to err on the side of caution, and an autonomous ship may therefore behave more slowly and more conservatively than an experienced master.

Communication adds another layer of risk. Remote control is only as good as the link between ship and shore, and satellite connections can be interrupted, delayed or jammed. Any vessel designed to be supervised from a control centre must therefore be able to cope with losing contact altogether, which means that its autonomy has to be genuine rather than nominal. Security specialists add a further warning: a ship steered by software and satellite is a target. A vessel that depends on data links offers an attacker more entry points than one steered by a wheel.

Trials have multiplied. In June 2022 an uncrewed trimaran called the Mayflower completed a crossing of the Atlantic from Plymouth to Halifax, powered partly by solar panels and navigated by software that had to identify and avoid other traffic. Several countries have tested autonomous ferries on short crossings, and harbour tugs and survey vessels are being adapted in the same way. Most of these craft operate in sheltered or well-mapped water, and nearly all are small. The long ocean voyage with no crew at all remains a demonstration rather than a commercial proposition.

Regulation is catching up slowly. The International Maritime Organization, the United Nations agency responsible for shipping, has identified four degrees of autonomy, ranging from ships with automated systems and seafarers on board to fully autonomous vessels, and is working on a code to cover them. Because existing conventions assume a master and a crew, questions of responsibility must be settled before insurers will underwrite a voyage: if a remotely controlled ship collides with a fishing boat, liability may lie with the owner, the software supplier, the sensor manufacturer or the operator in the control centre.

Ports remain stubbornly human. A large ship approaching a berth takes a pilot on board, is assisted by tugs and is guided by traffic control, and none of these practices adapts easily to a vessel with nobody aboard to handle a rope. Claims about the imminent arrival of crewless cargo shipping therefore tend to underestimate the shore-side infrastructure that would have to change with it.

The employment consequences are uncertain rather than trivial. The international fleet employs roughly two million seafarers, many of them from the Philippines, India, China and eastern Europe, and their remittances support whole regions. Automation would not so much eliminate these jobs as move some of them ashore, into control rooms staffed by people with different skills and, in most cases, a different country of residence. On present evidence the shift will be gradual, but it will be felt.

At half past nine on a Tuesday evening, a message arrives on a teacher's phone: a question from a parent, or a note from a manager about the next morning's meeting. The teacher replies, because not replying feels rude and because the answer will be wanted at eight o'clock. Multiply that moment by several hundred million office workers and the result is a change in the shape of the working day. For most of the twentieth century geography enforced the boundary between work and the rest of life: when you left the office, the tools of your trade stayed behind. Portable computers and smartphones removed that boundary, and with it the natural end of the day.

The laws that have appeared in response attempt to rebuild it. The phrase 'right to disconnect' is slightly misleading, since almost none of this legislation bans employers from sending a message. What the French version, introduced by a labour law of 2016 and in force from the following January, does is oblige larger companies to negotiate the terms on which staff must be available outside working hours. Portugal went further in 2021, prohibiting out-of-hours contact and attaching penalties; Ontario required employers above a certain size to publish a policy; and Australia adopted a similar right in 2024. The European Parliament has called for legislation across the Union.

The case for such rules rests on a growing body of evidence. Surveys in Europe and North America repeatedly find that employees expected to answer messages in the evening report more fatigue, poorer sleep and greater difficulty recovering from work, and the effect is strongest among those who feel they have no choice. The evidence is not as solid as it looks. Most of it comes from questionnaires in which people describe their own experience, and it is hard to separate the harm done by constant contact from the harm done by the work itself.

Opponents raise a different objection, and it deserves more attention than it usually receives. Many employees value the ability to work asynchronously — to leave early for a child's appointment and finish a task after dinner, or to answer a message from another time zone rather than wait for the morning. A law that treated all contact outside fixed hours as an intrusion would remove that freedom, and it would suit those whose lives are already orderly far better than those whose are not. The flexibility argument is not an excuse invented by employers.

The French experience also shows how much of this legislation is procedural. Companies were required to negotiate, but the law did not prescribe the outcome, and enforcement has been left to inspectors with many other duties. Where penalties are sharper, as in Portugal, the practical obstacle lies elsewhere: an employee who complains about a late-night message must go on working with the person who sent it, and few are willing to pay that price. Rights that cannot be used are, in effect, statements of aspiration.

There is a deeper problem still. The discomfort of constant availability is usually described as a problem of technology, or of manners, when it is often a problem of staffing. If a hospital ward is short of nurses, someone will answer the phone; if a contract specifies work that cannot be done within the contracted hours, no clause about disconnection will change the arithmetic. It is telling that the loudest complaints come from professional employees whose emails follow them home, while a care worker on an unpredictable rota, or a courier whose route and speed are set by an application, has no evening boundary to defend.

None of this means the reform is worthless. Its value lies in the default. At present the assumption in many workplaces is that a message sent is a message that should be answered, and that a failure to reply is a failure of commitment. A legal right, even a weak one, reverses that assumption and gives it a name. It also gives employees and their representatives something to bargain over, and it protects the junior member of staff who feels obliged to prove enthusiasm at eleven at night. Availability becomes a matter for negotiation rather than a matter of individual goodwill.

For such a right to amount to more than a signal of good intentions, it would need certain features. It should cover every kind of worker, including those on temporary contracts; it should specify in each contract the hours during which contact may be expected; and it should provide simple sanctions that an individual can trigger without exposing themselves to retaliation. The culture of an organisation is shaped, above all, by the habits of its managers: a supervisor who sends messages at midnight teaches the whole team that answering is expected. Training and example may achieve more than a clause in a statute, but a clause is what makes the training credible.

A right to disconnect is therefore worth having, not because it will cure overwork but because it draws a line where there is now nothing. It cannot by itself reduce the workload that makes people answer their messages at midnight, and it will achieve little unless it is enforced and unless it reaches the workers who need it most. What it can do is acknowledge that the working day has an end, and that the technology which made that end invisible can also be designed to restore it.