The reasons why air traffic controllers cannot be replaced by AI
Following a series of regular operations, an air traffic controller receives a radio communication from a small aircraft, expressing uncertainty about the extension of its landing gear for landing. In response, the controller coordinates with the pilot to fly at a lower altitude near the tower, allowing the controller to visually inspect the landing gear. Upon inspection, everything seems to be in order. The controller then informs the pilot, “Based on my observation, it appears that your landing gear is properly extended.”
The air traffic controller calls the airport’s fire engines to be ready just in case, and the plane circles back to land safely. Such scenarios are repeated regularly. In the air traffic control system, everything must meet the highest safety standards, but not everything goes according to plan.
Contrast this further with the science fiction of future AI “pilots” flying autonomous airplanes with an autonomous air traffic control system that handles the airplanes as easily as routers move data packets on the Internet.
I’m an aeronautical engineer who led a congressionally commissioned National Academies study of air traffic controller personnel. Researchers are constantly working on new technologies that automate elements of the air traffic control system, but the technology can only perform the functions that were designed during its design and thus cannot change standard procedures. As the above scenario shows, people are likely to remain an essential and central part of air traffic control for a long time to come.
What air traffic controllers do
The Federal Aviation Administration’s basic guidelines on the responsibilities of air traffic controllers state: “The primary purpose of the air traffic control system is to prevent the collision of aircraft.” Air traffic controllers are also tasked with providing “safe, regular and expeditious air traffic flow” and other safety-supporting services, such as helping pilots avoid mountains and other hazardous terrain and hazardous weather whenever possible.
The jobs of air traffic controllers vary. Turret controllers provide local control that clears aircraft for takeoff and landing, ensuring they are safely separated from each other. They also provide ground control, direct aircraft to taxis and notify pilots of flight plans and potential safety issues the day before the flight. Tower pilots are assisted by some monitors, but mostly look out of the tower and talk to the pilots via radio. At larger airports staffed by FAA air traffic controllers, surface surveillance displays show air traffic controllers the aircraft and other vehicles on the ground at the airport.
Approach and travel controllers, on the other hand, sit in front of large screens in dark and quiet rooms. They communicate with the pilots via radio. Their screens show aircraft locations on a map view, key features of airspace boundaries and routes.
The 21 air traffic control centers located in the United States control traffic between and above airports and thus typically fly at higher speeds and altitudes.
Air traffic controllers in approach control rooms transition from local control of departing aircraft after takeoff and into en route airspace. Similarly, they pick up incoming aircraft from en route airspace, position them for landing approach and hand them over to air traffic controllers.
A controller in each screen controls all intra-sector traffic. Sectors can vary in size from a few cubic meters focused on sequencing aircraft landing at busy airports, to route sectors of over 30,000 cubic meters (125,045 cubic kilometres) where and when there are few aircraft. If a sector is busy, a second and even a third controller might help, or the sector can be split into two parts, with one display and controller team controlling the other.
How technology can help
Air traffic controllers’ work is stressful and they have fatigue and information overload. Public concern over the increase in short-distance calls has put the spotlight on aging technology and staffing shortages, which have led to mandatory overtime for air traffic controllers. New technologies can help alleviate these problems.
The air traffic control system incorporates new technologies in several ways. The FAA’s NextGen air traffic systems initiative provides air traffic controllers with more — and more accurate — information.
The controllers’ screens initially only showed radar tracking. They can now utilize all the information known about each flight in the trip automation modernization system. This system integrates radar, automatic position reports from aircraft via automatic dependent surveillance transmission, weather reports, flight plans and flight histories.
The systems help alert air traffic controllers of potential conflicts between aircraft or aircraft that are too close to high ground or structures, and provide air traffic controllers with suggestions on how to organize aircraft into smooth traffic flows. In testimony to the US Senate on November 9, 2023 on airport security, FAA Chief Operating Officer Timothy Arel said the administration is developing or improving several air traffic control systems.
Researchers use machine learning to analyze and predict aspects of air traffic and air traffic control, including intercity air traffic and air traffic controller behavior.
How technology can complicate things
New technology can also cause profound changes in air traffic control in the form of new types of aircraft. For example, current regulations mostly restrict unmanned aircraft from flying below 400 feet (122 meters) above the ground and away from airports. These drones are used by first responders, news organizations, surveyors, delivery services and hobbyists.
However, some emerging UAVs propose flying in controlled airspace. Some plan to have their aircraft fly regular flight routes and normally interact with air traffic controllers via voice radio. These include Reliable Robotics and Xwing, which are separately working to automate the Cessna Caravan, a small cargo plane.
Others aim at new business models, such as advanced air transport, the concept of small, highly automated electric aircraft – for example, electric air taxis. These would require dramatically different air traffic routes and procedures.
Expect the unexpected
An aircraft that requires special handling can disrupt the air traffic controller’s routines. This can range from an emergency to priority handling of medical flights or Air Force One. Controllers are given the responsibility and flexibility to adapt the way they manage their airspace.
The requirements of air traffic control’s front line do not match the capabilities of artificial intelligence. People expect air travel to continue to be the safest, most complex, high-tech system ever. It achieves this standard by following procedures when practically possible, which is what AI can do, and by adapting and using good judgment whenever something unplanned happens or a new operation is executed – a notable weakness of today’s AI.
In fact, when conditions are at their worst—when air traffic controllers are figuring out how to handle aircraft with serious problems, airport crises, or general airspace closures due to safety concerns or infrastructure failures—air traffic controllers make the greatest contribution to safety.
Air traffic controllers don’t fly the plane either. They communicate and interact with others to control the aircraft, so their duty is essentially to work as part of a team – another notable weakness of AI.
As an engineer and designer, I am very excited about the possibilities of artificial intelligence to analyze big data recordings of previous air traffic operations, for example to achieve more efficient flight routes. However, as a pilot, I am happy to hear the calm voice of the air traffic controller on the radio to help me land quickly and safely if I have a problem.
