Forget standing seats, in-flight VR or luggage-tracking apps. It’s unmanned flight that could be aviation’s next big transformation.
At this year’s Paris Air Show, Airbus said it’s trying to attract aviation regulators to the idea of pilotless commercial travel. So is rival Boeing.
Their timing couldn’t be better. With demand for air travel soaring, over 800,000 new pilots may be needed over the next 20 years. However, the supply of new pilots is struggling to keep up with demand, producing what Boeing has called “one of the biggest challenges” facing the airline industry.
But while pilotless technology offers relief, it poses challenges of its own that could ultimately stand in the way of autonomous airlines taking to the skies. Here are three of them.
Innovation invariably creates winners and losers. The introduction of the automobile shifted consumer demand away from trains much like the railways had, in decades prior, displaced canals and waterways as major forms of transportation. The result was job offers for some workers and pink slips for others. This reality is best summed up by Nicholas Carr, in his book, The Glass Cage, Automation and Us: “There is no economic law that says that everyone, or even most people, automatically benefit from technological progress.”
You might also like:
- Norway’s plan for a fleet of electric planes
- The plane that shrank the world
- How flying messes with your mind
Pilotless planes are a prime example of this. While the technology promises to revolutionise travel, its asking price is jobs – specifically, piloting jobs. The airline industry employs tens of thousands of aviators worldwide – skilled professionals who ferry billions of passengers across trillions of kilometers. Delegating this task to machines would produce widespread unemployment among pilots, culminating in a struggle to ply their skills to a new trade. That's hardly an easy task considering the unique skillset flying demands.
That’s where politics come in. Airline pilots are backed by powerful labour unions, organisations that use collective bargaining, campaign contributions and political lobbying to influence issues affecting their members.
American Airlines recently spent $25bn (£20bn) upgrading its fleet (Credit: Getty Images)
Take the Air Line Pilots Association (Alpa). Representing over 63,000 aviators worldwide, a compelling example of Alpa’s influence dates back to the 1960s. In decades prior, airplanes required a third crew member in the cockpit. This flight engineer monitored airplane instruments and assisted pilots with troubleshooting. However, technological advances made flight engineers obsolete and manufacturers started producing airplanes with only two crew in mind.
However, given the inevitability of job losses among its members, Alpa resisted the adoption of these craft, waging what was by one account "a long struggle, contractually to get management to give (flight engineers) meaningful duties”. Similar tactics are likely should pilotless technology reach maturity. Alpa has already voiced opposition to further reducing crew numbers.
Though crashes are rare, they do happen and when they do, airline balance sheets take a hit
Labour unions have allies in their fight against algorithms. That’s because automating flying doesn’t just threaten commercial airline jobs. Corporate flight academies train an increasing number of today’s aviators. So do universities, colleges and small flight schools. These institutions employ tens of thousands of personnel – like flight instructors, ground lecturers and simulation experts – to run pilot training programs worldwide. Pilotless technology also threatens these professions.
Aircraft aren’t cheap. The Boeing 737 – a small, single-aisle passenger jet – costs upwards of $100m (£80m) apiece. Its larger, twin-aisle counterpart – the Boeing 777 – costs over $300m (£240m). Though airlines get discounts for bulk orders, the final bill can still be substantial. In 2011, American Airlines spent over $30bn (£24bn) on revamping its fleet. Indigo Airlines – an Indian start-up carrier – forked out an equally hefty amount for its fleet. Recovering those costs means putting airplanes to work and that can be risky. Though crashes are rare, they do happen and when they do, airline balance sheets take a hit.
That’s where insurance comes in. Airlines buy policies to cover losses if an aircraft is damaged or destroyed. These policies also protect carriers against claims stemming from crash-related injuries and property damage. Exactly how much airlines pay for insurance coverage is anyone’s guess – insurers guard their rates to maintain a competitive edge. However, premiums collectively cost the industry billions annually, making this expense a key determinant of an airline’s potential profit (or loss).
Airline pilots are backed by powerful labour unions and are resistant to mass lay-offs (Credit: Getty Images)
An important question facing pilotless technology is how its introduction will affect premiums. Conventional wisdom suggests premiums should drop. After all, pilotless planes – powered by a sophisticated array of sensors and software – eliminates the culprit implicated in most airplane accidents: humans. This means that while crashes today are rare, they will – in the era of pilotless flight – be rarer. Fewer crashes will mean fewer payouts by insurers and the ensuing savings should be passed along to airlines in the form of lower premiums.
At least, that’s the idea. The reality may be more complicated.
Today’s airliners are already algorithmic heavyweights. Unlike their predecessors that relied on mental muscle to fly, modern jets count on software. The Boeing 787 Dreamliner – a staple of fleets worldwide and state-of-the-art aircraft – is powered by millions of lines of code. More code allows for more safety-enhancing features – things like turbulence detectors that pre-emptively sense rough air, diagnostic systems that monitor an aircraft’s health and smart sensors that automatically stabilise an airplane if the cabin depressurises.
But more code, while used in the interests of improving safety, also creates a new type of risk, replacing one type of human frailty for another. In 2015, government regulators warned that the Dreamliner could experience failures of its electrical system. The reason? A software glitch – undetected by engineers – that could lead to a “loss of control of the airplane”. Airbus recently faced similar issues with its flagship jet, the A350. The plane was found prone to a “partial or total loss of some avionics systems or functions, possibly resulting in an unsafe condition”. The culprit, once again, was software bugs that engineers failed to catch.
As software’s influence becomes absolute, the number (and danger) posed by coding follies will grow
A related algorithmic risk is hacking. As code becomes more complex, thoroughly testing it for weaknesses becomes harder. This opens the door to exploitation by hackers. In 2008, government regulators warned that the Dreamliner’s flight control system could be accessed via the airplane’s entertainment system, allowing passengers to override pilot commands.
According to a government document, the airplane’s design allowed for “new kinds of passenger connectivity to previously isolated data networks connected to systems that perform functions required for the safe operation of the airplane”. More recently, a team of professionals managed to remotely hack a Boeing 757 using radio waves.
The Boeing 787 is kept in the air thanks to millions of lines of code (Credit: Getty Images)
These risks are set to rise with pilotless planes. As software’s influence becomes absolute, the number (and danger) posed by coding follies will grow. This doesn’t mean the pilotless planes won’t be insured. There is a saying amongst underwriters: there’s no bad risk, only a bad rate.
Put simply, anything is insurable if the price is right. But what is that price and is it one airlines are willing, or are even able, to pay?
Pilots pocket much of an airline’s earnings. The average pay for flying a jetliner – assuming at least five years of experience – starts at $147,000 (£117,630) annually. Senior pilots can make nearly $300,000 (£240,000). Such hefty figures reflect the reality facing an industry keen to control costs: pilots are in short supply and learning to fly is both arduous and pricy. These factors push salaries up making them a key influencer of an airline’s labour cost and – along with fuel – among an airline’s biggest expenditures.
Pilotless technology should of course change this. Swiss bank UBS estimates that removing humans from the commercial cockpit could produce savings upwards of $35bn (£28bn) annually. That figure would boost profits in an industry that has often struggled to make money.
Algorithms’ cash-saving virtues are well-known. From energy to retail to defence, software is widely credited with slashing costs by automating tasks once delegated to humans. Why can’t the same approach be adopted in commercial aviation?
Airliners aren't cheap - and any accidents can cost the airlines which operate them many tens of millions of dollars (Credit: Getty Images)
Aerospace manufacturers seem to think it can. In 2017, a Boeing executive embraced the idea of pilotless airplanes, saying, “the basic building blocks of the technology are clearly available”. Similar sentiments were voiced by his Airbus counterpart earlier this year. The company’s chief commercial officer said Airbus already has “the technology for autonomous flying”.
Manufacturers could try curbing the number of crew on board
Yet autonomous does not mean humanless. While sensors and software curb the need for manual labor, that need isn’t – contrary to we’re told – purged entirely. In fact, you’d be hard pressed to find one industry where algorithms work without any human involvement or oversight whatsoever. The reason? Automation is imperfect. It errs just like humans do, and when that happens the results can – depending on the industry – be catastrophic.
Take the airplane autopilot. First introduced in 1912, “George”, as it is colloquially known, has become a staple of the modern cockpit. When engaged, its algorithms can crunch data faster and more reliably than a human pilot ever could, ultimately producing a safer, smoother flying experience.
But autopilot failures have also been implicated in several air incidents and crashes. That’s why its use is only approved under the watchful eye of a human pilot. Regulators know that George can’t – for all his virtual virtues – be trusted to get it right all the time, every time. You’d be hard pressed to find an aerospace manufacturer who disagrees with such reasoning.
If ground controllers help 'fly' planes, how many will be be expected to look after at the same time? (Credit: Getty Images)
Manufacturers could try curbing the number of crew on board. UBS estimates that going from two to one pilot in the cockpit would still produce some savings (though $20bn (£16bn) less than going fully pilotless). According to UBS analyst Celine Fornaro, so-called “single-pilot operations”, could become a reality as early as 2022.
But this proposition is equally problematic. That’s because it assumes the sole human pilot will intervene on time, every time automation fails. Manufacturers are wary. This explains why the single-pilot cockpits – being developed by the likes of Boeing and Airbus – includes remote supervision.
The idea is simple. With one pilot behind the controls, a second always stands ready to provide additional support. However, rather than being in the cockpit, the second aviator monitors the aircraft from the ground. It’s a nice idea, but it is one which raises an important question – how many airplanes should the second pilot watch? After all, paying this individual to just monitor one aircraft wipes out the $15bn (£12bn) cost advantage offered by a single-pilot cockpit. Under these conditions, airlines are better off keeping two aviators on the flight deck. But if remote monitoring costs can be distributed over multiple aircraft, the idea becomes economically more feasible.
Can aerospace manufacturers and airlines guarantee this setup – one that tests the limits of human attention and memory – won’t compromise passenger safety
It also becomes riskier. After all, can one remote pilot really be expected to help a distressed aircraft while keeping an eye on several others? What happens when multiple aircraft need help? Can aerospace manufacturers and airlines guarantee this setup – one that tests the limits of human attention and memory – won’t compromise passenger safety?
Until they can, the concept is unlikely to fly.
Ashley Nunes studies transportation safety, regulatory policy, and behavioral economics at MIT.
If you liked this story, sign up for the weekly bbc.com features newsletter, called “The Essential List”. A handpicked selection of stories from BBC Future, Culture, Worklife, and Travel, delivered to your inbox every Friday.