There are many publications crystal balling autonomous airliners, and preaching for fully autonomous flights. Big players seem to follow suit. FedEx and UPS signed a 134-million-dollar deal with a startup to design autonomous flight to tackle the pilot shortage and reduce costs. A few years ago, an investment bank predicted that pilotless aeroplane could become reality in 2025, and make the industry save up to $ 35 billion per year. On the other side, unions such as the Air Line Pilot Association (ALPA) are expressing great concerns, and believe pilots avert aviation disasters citing US Airways flight 1549.
The concept of autonomous flights has gained some sort of momentum lately, but all indicators point out that flight crews will remain. Here are two reasons why.
1. Humans Are Irreplaceable
In this fourth Industrial Revolution, technology is seamlessly integrated in our physical, digital, and biological worlds. One can safely claim that it has become the norm in our day-to-day tasks.
Aviation also thrives on tech. In fact, under certain caveat, tech has significantly enhanced safety and operations. It would be fair to say that the introduction of drones and remote piloted aircraft systems (RPAS) to the market in the mid-2000s for commercial purposes was a decisive moment in aviation history. This commercialization accelerated into the mid-2010s with the release of popular models which incorporated GPS and cameras making the technology accessible to consumers. On the regulatory side, ICAO adopted in 2024 new SARPs to enhance safety and integration of RPAS into the global air navigation system. ICAO amended the Annexes to the Chicago Convention, and established requirements for remote‑pilot licences, operator certificates. ICAO is currently working to integrate RPAS into non-segregated airspace (i.e. airspace shared with manned aircraft) safely.
a. RTK
Now, many believe autonomous aircraft will be the next logical stage of this technological development. Automation is not a foreign concept to aviation. For decades, aircraft manufacturers have used automation, the most obvious examples being autopilot systems (i.e. Automatic Flight Control System and Flight Management System), but recent developments in artificial intelligence and autonomous vehicle definitely pushed the fully autonomous flights.
Real-time kinematic (RTK) technologies are probably today’s best bet. RTK is a GPS correction method that enhances precision from standard meter-level to centimeter-level accuracy (1 to 3 cm) by using a ground-based reference station. RTK is instrumental in making drone location data more accurate. RTK corrections are essentially derived by calculating the positional discrepancy between a nearby base station and the drone and result in centimeter-level accuracy positioning. Thanks to RTK, one can contemplate aerial display of pre-programmed flocks of drones equipped with LEDs swarming and changing patterns. RTK is also used in different aerial work applications such as in the agricultural and surveying sectors.
As it concerns transportation, RTK could be used in eVTOL and Urban Air Mobility (UAM) vehicles. The level of precision (1 to 3 cm) would be suitable to conduct precision takeoff and landing and geofencing. That being said, this tech cannot be the sole navigation method due to its dependence on satellite signals that could be blocked by buildings, terrain, bad weather, etc., and lack of redundancy to carry passengers.
b. Decision-making
Another key aspect limiting the use of fully automatous for passengers’ concerns decision-making. Decision-making is paramount in aviation. Pilots learn to use this skill at their first lesson with their flight instructor. The human brain is a bioengineering system by itself with thousands of years of development. Our brain perceives dangers, threats and emergencies. It can also learn how to address these perils.
Pilots are professionals who complete stringent training throughout their career. From pre-flight planning to disembarkation, they constantly evaluate information and make critical choices. This is why flight experience is absolutely vital. Pilots must possess strong judgment, situational awareness, and problem-solving skills to handle unexpected events and ensure the safety of the aircraft and everyone on board.
In 2016, the Department of Transportation’s Inspector General due to several accidents, including the July 2013 crash of Asiana Airlines flight 214, published a damning report warning airlines to better train their pilots with their manual flying skills. The FAA stated that “reliance on automation is a growing concern among industry experts”.
AI is data-driven and probabilistic. It makes decisions based on patterns and probabilities from their training data, rather than on logical, deterministic rules. AI is a sophisticated calculator. It has no intuition, no experience especially when data is missing. Meanwhile, humans rely on intuition and past experiences to inform their decisions especially when information is incomplete. Moreover, in critical situations, autonomous systems may face ethical dilemmas that are difficult to program effectively or justify legally.
While AI will probably be used to assist airline pilots by reducing workload and at some extent in their decision-making processes, it will not replace them.
c. Legal issues
Although progress is significant in regards to RPAS as discussed above, ICAO has not yet any standard or regulation regarding fully autonomous aircraft. ICAO still has to define what constitutes “autonomous aircraft” vs high automation vs remotely piloted, in terms of regulatory requirements, and to identify and plan the certification of AI systems (machine learning models, behaviour under edge cases, explainability, accountability, validation) are still emerging.
Another potential legal issue concerns liabilities. Who would be held liable if an accident of a fully autonomous aircraft transporting 300 passengers occurred? Would the AI system have its own legal personality, would the AI manufacturer be entirely held liable? Also, under domestic law, if the pilot-in-command is defined as a peace officer, in the context of an AI aircraft, who would be the peace officer to ensure the safety of the flight and passengers aboard? Who airlines need to hire air marshals? Would this air marshal know how to operate the aircraft?
2. Public Trust
Low public trust is perhaps the most conclusive argument against autonomous airliners. UBS performed a survey with 8,000 people in the US, Europe and Australia. It found that only 17% percent of travelers are willing to fly without a pilot. Meanwhile, 54% said they would be unlikely to buy a ticket for such a flight. Most passengers are reluctant to trust a completely autonomous aircraft without a human pilot onboard, especially during emergencies. Even if statistically safe, the perceived lack of control can significantly impact consumer acceptance, reducing demand. How would airlines survive if more than half of passengers would not buy a ticket for an automated flight?
ALPA retained IPSOS to conduct a recent online study. It was revealed that Americans are much less likely to feel safe when it comes to automated air transportation. Only 15% say that they would be comfortable as an airline passenger on an airplane that was completely pilotless. 81% claim that they would be not very/not at all comfortable with this. Americans are most likely to feel rail transportation as a good fit for automation (82%, rated 1,2). Just over 50% believe that water transportation (e.g., ferries, cruise ships) is best suited for automation technology to allow a “self-driving” mode to carry passengers (57%), while 43% say the same thing of road transportation. Americans are least likely to believe that air transportation is best suited to be equipped with such automation technology (19%).
More than 80% Americans believe that two pilots working together is the best option when it comes to solving problems, troubleshooting system issues, and/or making split-second decisions during a flight/when dealing with an emergency. Only 8% think that a single pilot is best suited to deal with these types of situations, while fewer believe that an on-board computer (7%) or remote flying technology (5%) would be best for solving problems/trouble shooting system issues.
Numbers speak for themselves.
Conclusion
AI offers great solutions to commercial aviation such offering: (1) maintenance issues predictions, (2) flights paths, fuel usage and scheduling optimization, and (3) passenger experience and streamlines airport logistics improvements. There are even companies working on the development of an AI-powered pilotless C208.
It’s probably fair however to claim that we are currently in an AI bubble apparent to the dotcom bubble. A study published by MIT in 2025 stated that 95% of generative AI pilots at companies are failing.
To conclude, humans will keep flying aeroplanes, especially airliners.
