I would like to know what pilots are thinking about this subject. I am 16 years old and I have always wanted to become a pilot. I have plans to start my pilot training after upper secondary school. The price of the pilot training does not scare me at the moment there is something else holding me back. I have noticed that Airbus and other plane manufacturers have started research into automatic planes (without pilots). So that is why I am thinking if it is worth becoming a pilot for me. I could potentially start my training in 3 years time and be graduating 5 years from now. Then I would hope to have a career in aviation before I retire. I have heard that there is speculation that within 20 years from now there could be passenger planes without pilots.
I can see at the moment that the planes that are being produces still require pilots and I know these planes will continue to fly for a number of years but I am interested in hearing what you think is going to happen?
Thank you for the question! Your concern is something that a lot of people wonder at the moment and we get many questions regarding the subject. Will there be a time when pilots are not needed? To be honest we don’t have a crystal ball but we do know how the industry generally works and we will share our view about the topic.
We have heard rumours that there’s already now a possibility to make a civil plane without pilots flying the plane. If we look into the military world there are also numorous examples of the use of large drones. But if there’s already technology to produce civil planes that can fly without pilots, why are they not out already? We think this is mainly down to the regulatory process. The aviation world can be really stiff and it will take a long period of time before plane manufactures can get approvals for new technology. It might be even harder to get approval after what we have experienced with the B737 MAX. The B737 MAX perhaps serves a reminder of how automatisation/ technology can sometimes also lead to problems.
Here are some of the challenges that we can think of when it comes to having a fully automatic flown aeroplane with no pilots onboard.
Routings and re-routings
Even if the automatic aeroplane would always fly between the same point of departure and to the same destination, the route between the two is likely to be different almost every time. Flying is not like going on a rail where the route is pre-determined by the rail. We don’t have rails up in the sky. Also weather like thunderstorms can affect the routing and this can also change mid-flight last minute. It would be important for the operator that the automatic aeroplane could be used just as flexible as an aeroplane with pilots. This means a lot of routes and airports would have to be certified for automatic aeroplanes. This alone could be a huge certification task. Of course the routing could be fed to the computer by operators on the ground but how would the computer cope with last minute changes of the runway or a last minute change of the routing in-flight?
Weather is something that can sometimes be seen from the flight deck with the human eye but not by a weather radar. Let’s say that ATC instructed the aeroplane to turn left and the pilots decline it because they can see with their eyes that there’s a thunderstorm that the weather radar doesn’t pick up – it could be for example the tip of the “anvil” of a thunderstorm. The question is if the computer is able to make these kind of assessments? Before you could safely remove the pilots from the flight deck, we think weather radars would have to become more reliable and sophisticated than what we see today. Weather is also something that is constantly changing. As a pilot you can listen to the frequency and for example hear other pilots that are reporting clear air turbulence (which cannot be detected by the weather radars we have today). By listening in on the frequency you can also make decisions based on what other pilots (outside the range of your weather radar) are doing or experiencing. Will a computer also be able to make those kind of decisions based on reports on the frequency made by other pilots?
Let’s say that there’s a medical situation on board. Usually the decision to land is ultimately made by the captain but in the decision making process Medlink is often included. Medlink is a team of medical professionals on the ground that the pilot can contact for medical advice. Medlink has knowledge about what tools and medication the crew have on board the aeroplane and Medlink can make an assessment of what kind of treatment the passenger needs. This is however based on the kind of information that the pilots provide Medlink with. As the pilots are on the flight deck and the passenger is in the cabin, the cabin crew are involved in the communication loop. Together with cabin crew, pilots and Medlink will come up with a plan that is best for the passenger. When deciding where to land the facilities of the airport (and adjacent cities) will often be considered. There’s no point in taking a passenger to a hospital that doesn’t have the needed tools or doctors for the treatment. This is where the knowledge that Medlink has is useful but also the pilots need to consider what range of airports they can safely land at. Of course there needs to be communication between pilots, cabin crew and Medlink. Cabin crew will monitor the situation and will inform the pilots on the condition of the passenger and inform if there are any changes in condition. Once again this is information that can affect the decision making (to land or not to land) so it is important that the information is given timely and precise. In a totally automated aeroplane it would be interesting to see how they solve the 3 way communication in these situations as it can put the life of a passenger on the line. Can a computer make a decision and communicate with Medlink and Cabin crew in the same way pilots do today?
Communication with ATC (air traffic control)
Communication is something that is really hard. Even though the spoken language in Aviation is English, there are so many accents of English that the computer would need to learn if any system is based on voice recognition. Already now some airspace have CPDLC (Controller Pilot Data Link Communication) which is where the controller on the ground can communicate and provide instructions to the pilots without voice. The pilots receives the instruction(s) and can approve the instruction(s). This way the pilots are still in control. It is likely that automatic aeroplanes would be restricted to areas with CPDLC or similar. At the moment that would mean that there would be many areas of the world where they could not fly. It could be a significant challenge in the industry to build the required ATC technology required to support automatic aeroplanes and we can see how the “who will pay” will be an issue that could halt introduction.
Also some of the decisions that pilots make when they fly are based on listening to what is happening on the frequency. This could for example be a decision relating to speed, altitude or routing. For example if ATC have asked multiple aeroplanes to slow down and you can see that you are catching up and next in line, you could make the assumption to slow down before the controller asks you or you could maybe at least anticipate it. Also sometimes you can have a conversations with ATC that can improve your routing (directs) and this can save both time and money. Pilots will also regularly negotiate different routings and flying levels that will help another pilot out. It would be interesting to see how Airmanship and communication with ATC would be solved in aeroplanes with no pilots.
Will it be more cost effective?
An important reason to get automatic aeroplanes would be saving money. So the automatic aeroplane has to be cheaper and more effective than two-pilot aeroplanes. The first obvious savings would be two pilot salaries. Of course a computer doesn’t care what time of the day it is or how much rest there is between flights and therefore it can be used more often. While the automatic aeroplanes would make the pilot role, as we know it today, redundant there would probably still be the need for engineers. But will the automatic aeroplane require new roles? For example somekind of aeroplane supervisor on the ground that could take over? We don’t know at the moment of course but if an automatic aeroplane would be more effective (flying more) there could be a need to have more engineers doing frequent checks and it could cause more engineering expenses. Not only work hours but also if new technology or training courses was required to service these type of new automatic aeroplanes. Also our experience is that computer systems don’t like power failures or situations where they need to reboot. On one occasion we experienced shutting down a 787 Dreamliner completely and it took over 3 hours of engineering work to get it up and running again. It was like rebooting a computer and then having multiple pop ups from systems that needed some kind of interaction in order to work. We wonder how long it would take if you had something like that with a 100% automated aeroplane.
You may feel that too many times we have seen that the reason for an accident or a crash is a human error. Sadly accidents are down to human error in most of the cases. What is important to state though is that it is sometimes a technical error that then is followed by a human error. The question is if a computer is a safer option? It would of course remove the risk of human error but what about multiple technical errors or a failure of the computer systems controlling the aeroplane? There are two pilots on the flight deck for very good reasons. One is to prevent human error by having another human in the loop. Computers might be better at making safer choices based on black and white facts or algorithms but what if there is a computer error or an error in the inputs to the computer? How can a fully automated plane with no pilots override it’s own errors? What often makes our industry safe is the professional scepticism and experience of the pilots. Errors generated by electronic systems are often caught by pilots – for example load sheet errors. Pilots constantly compare what they see “outside” the aeroplane with indications from the “inside” or what they expect with what they have. Handling a technical malfunction such as “unreliable airspeed” is basically back to basics – attitude and power. There are also a number of technical scenarios where the safe outcome is ensured by the pilot disengaging the automatics and taking control. How would the computer be able to do that? Once again we can use the 737 MAX as an example of where a computer system was supposed to make an aeroplane safer but how technology ended up creating a problem instead. How would a computer be able to cope with those situations where there is a failure in the data it is provided with? How will it be able to do the same things that pilots are able to do today where they can take control if something is not right?
So in the end we would like to question if an automatic aeroplane with no pilots can be made so it is both cheaper and more safe than having two pilots? At the moment it looks like developing a computer that can make totally independant decisions, communicate with crew, consider all variables while flying – might be more expensive than having two pilots. As long as pilots are cheaper and they can come up with better solutions using the “human factor”, we cannot see automatic planes being a big threat to your career. In some cases computers could definitely make better and safer decisions than pilots but the remaining big questions is still how can it over ride itself? On the road to an aeroplane with no pilots we would probably see an aeroplane with just one pilot. From there it would probably be many years before that one pilot is removed from the flight deck. But if and when you start seeing a change and big commercial passenger aeroplanes start to fly around with just one pilot, then we would start worrying about the future for civil pilots. At the moment there is no indication of that happening anytime soon so while we of course do not have a crystal ball we cannot see automatic planes being a big career threat to someone who starts their pilot training in the next few years.
For example Airbus has done pilotless take-offs and you can read more about it HERE.
You mention that it could be 3 years before you start your training and a lot can change in 3 years, so perhaps now the best you could do is to follow what is happening in our industry and have a plan B, just in case you decide not to become a pilot. For someone who thinks about starting pilot training now we suggest reading our blog about: Is it now good time to start pilot training?
Also if you’re still really unsure about the pilot career, we recommend to do a lot of re-search. We have also written an E-book about what it is like to be a pilot and how to choose pilot school etc. You can read more about it HERE.
ASK A PILOT