In June of 2017, Boeing announced that it was testing autonomous aircrafts, with AI making the decisions that pilots are currently entrusted with. In 2016, a team of engineers unveiled ‘Pepper’ an Inflight Service Bot who could pick up garbage, circulate the plane during turbulence, and understand passengers speaking numerous languages. Even as early as 2012, a company called Skymax had built a ‘Skytender’ — or robotic bartender to assist the flight attendants. The travel industry has always been tech savvy — adopting chatbots, mobile apps, and automation well before other industries. Is the next step replacing humans with AI?
Why Replacing Humans With AI Does Not Make Fiscal Sense for Commercial Flights
The future of commercial airlines may be AI, but if it is, it won’t be because of financial reasons — making it a unique industry. Unlike grocery store cashiers, automobile drivers, or even doctors and lawyers, replacing pilots with AI would actually be a cost-ineffective measure.
According to NewtonX experts who have held senior level positions at Uber Elevate, Waymo, and major commercial airlines, the cost of hiring pilots and flight attendants is miniscule in comparison to the cost of building out a plane. The reason there’s no parallel between the benefits of autonomous vehicles for, say, Uber, and the benefits for commercial airlines is a simple computation of the cost of the machine vs. the cost of the human running the machine.
The total cost of ownership for a car is roughly $9,000. When you employ a driver for a car, the cost increases between $25,000 and $40,000 because of the cost of employment. For commercial planes, however, the total cost of ownership is roughly $50M. Even if you add up the cost of all the people associated with that airplane (the pilots and cabin crew), the maximum total added cost is $1M. Investing in AI to replace humans would diminish the total cost by less than 4% — and that’s a generous estimate. To boot, changing plane equipment costs millions, not to mention how much investment in research and the development of autonomous technology costs.
The Goal of Autonomous Aviation is not Cost Savings; it’s Safety
Much of aviation is in fact already automated — after all, it’s where the term autopilot came from. Flight management systems do 95% of the work associated with manning the plane, according to a VP-level executive with United Airlines. In fact, the role of the flight engineer was eliminated in the 1980’s after controls were automated and placed in the pilot’s panel. Redundant pilots are only used today on long flights to ensure that all crew members can get enough sleep.
Airplanes are currently controlled by a system called fly-by-wire that takes care of setting courses, switching radio frequencies, and entering data into flight systems via sensors. These automated systems have dramatically improved flight safety — they were introduced beginning in the late ‘80’s and since then airliner fatalities have steadily decreased. 2017 was the safest year in history with only 10 fatal airplane accidents, resulting in 44 deaths (worldwide). In the United States between 2000 and 2010 there were only .2 deaths per every 10 billion passenger-miles. Compare this to 1929 when there were 51 total commercial airline crashes.
From predictive maintenance analysis to in-flight performance characteristics, AI and automation have made flying one of the safest forms of transportation. And increasing sensors and technological redundancy will only make flights safer. In fact, even robotic flight attendants are intended to maximize passenger comfort while minimizing risk to the cabin crew — not to eliminate human jobs to maximize efficiency. The decision making abilities of robotic flight attendants aren’t advanced enough to actually replace humans, but they can roam the aisles during turbulence without fear of falling, and can improve sanitation by serving food and drinks with sterilized robotic arms.
Automation can lead to complacency for the humans it aides, though. There are concerns over the ability of pilots and cabin crew to effectively take over in the case of disaster. For instance, the 2009 Air France Flight 447 crash, which resulted in the deaths of 228 passengers and cabin crew, occurred because the autopilot disconnected and the pilot and crew reacted to it incorrectly, causing the plane to go into aerodynamic stall. Incidents like this highlight the importance of having well-trained pilots who understand how to act as safety redundancies when technology fails.
The Market Opportunity For Personal Self-Flying Vehicles
While the true value in increased automation for commercial aviation is safety, there is a massive potential market opportunity for 1-4 person autonomous flying vehicles. This is a different market opportunity, one that is already being explored by Audi, Uber Elevate, and Airbus. In fact, Dubai has already begun testing autonomous flying 2-person taxis, called Autonomous Air Taxis (AAT), and the government of Dubai aims to have autonomous transport account for a quarter of total trips by 2030.
These ‘flying cars’ have myriad benefits, including their low emissions. Because they use electric propulsion, and are powered by battery, they provide a clean alternative to traditional, fuel-based cars. That said, the energy per unit weight provided by battery is still insufficient for long commutes — thus if self-flying vehicles are launched before the battery is improved, they will need to have a mile cap on them. Batteries also still take time to charge between rides, which decreases the profitable hours that autonomous flying cars provide. There are promising battery advancements on the horizon, however, and considering that batteries are a primary area of focus for autonomous electric vehicles, it’s likely that there will be major breakthroughs in the next few years.
Next Steps for Autonomous Air Travel
Autonomous commercial flights with a single backup pilot will be the norm in the next 21 months, according to NewtonX experts. Autonomous flying cars, however, will take longer to be implemented because of the red tape around urban air travel, as well as consumer hesitance over autonomous vehicles. Before flying taxis are allowed in major U.S. cities, we will see drone delivery and autonomous cars on the ground — all of which will take at least another decade to fully implement.
While heavy automation in the skies has already proven to improve safety, incidents such as Flight 447 ensure that there will always be a level of human control. Even if fully automated commercial planes are developed, pilots will still be needed as a form of backup failsafe because of the sheer number of human lives at stake.