Airline safety: 3 ways AI can help make flying even safer

A string of plane crashes this year, including the catastrophe over the Potomac River near Washington D.C. rightly thrust airline safety into an unwelcome spotlight, one that always brings up the question of how safe it is to fly commercial.

The short answer is, it’s safest form of transportation out there. But as recent events have shown, it can and must be safer, and advances in three areas of aviation technology—navigation and flight control, predictive maintenance, and air traffic management— can lead the way. Artificial intelligence figures heavily into all three.

First, some safety data to settle the nerves. The most recent International Air Transport Association (IATA) annual report on safety among 340 airlines in more than 120 countries found that the five-year average of airline accidents per “sector”—aviation-speak for a flight leg involving a takeoff and a landing—was 0.16 accidents per million.

So, if you were to fly commercial every day, it would take, on average, more than 19,000 years to end up in a fatal plane crash. As aviation-safety expert Anthony Brickhouse told CNN, “statistically speaking, you’re safer in your flight than you were driving in your car to the airport.”

That said, the airline business faces two big safety challenges that compound each other: increasingly crowded airspace and staffing shortages from cockpits to control towers to maintenance hangars. AI technologies promise to tackle both. Let’s start in the cockpits.

AI + airline safety: Navigation and flight control

Advanced navigation systems help keep pilots on course to reduce collision risk and optimize flight paths. Fortifying these systems with AI can help guide aircraft more precisely and without heavy reliance on ground-based equipment and error-prone manual input.

AI can take into account variables like weather patterns, air traffic, and fuel consumption, and even where pilots are focusing their attention to boost safety and efficiency at once.

AI also plays a big role in automated flight control technology. These systems can already handle takeoff, landing, and altitude maintenance in commercial jets. Efforts to advance automated flight control technology include two notable Airbus research programs:

1. Airbus’s Wayfinder is building scalable, certifiable autonomous systems to address flight-crew overload. The program aims to lessen pilot workload, boost safety through automated trajectory projection, and digitally assist collaboration between flight crews and air traffic control.
2. Airbus’s Optimate program is an autonomy demonstrator that aims to automate flights gate-to-gate with 4D trajectory flight management, virtual assistants, and overridable protections to support flight crew. Optimate, though, is starting on the ground, focusing first on increasingly chaotic taxiing.

Fully autonomous commercial air travel appears to be a distant destination, with an ETA pushing 2050, the European Aviation Safety Agency estimates. Military aviation looks to be moving much more quickly in that regard.

Beyond ‘Tuesday Booking’: How AI is revolutionizing travel with examples and benefits

AI is reshaping the global travel industry as is continues to recover from the pandemic. With AI, travel companies can surpass traveler expectations, boost efficiency, and streamline compliance.

Predictive maintenance and MRO

The IATA says nine of the 29 nonfatal commercial accidents reported in 2023 among its member airlines were related to landing gear. Predictive maintenance, a fast-growing airline-industry focus, could have made a difference in some of those—and in countless other instances where mechanical failures impacted safety, airline operations, or customer service.

Predictive maintenance takes historical maintenance records, IoT sensor data, and inputs from various aircraft systems to help spot, fix, or upgrade components before they fail. AI sifts through the data to generate maintenance-and-repair recommendations. Integration with cloud ERP software then ties maintenance tasks to supply chain, procurement, HR, service, and other affected business areas.

Of course, somebody must actually do the maintenance and repairs. Air-travel growth, aging fleets, a graying workforce, and maintenance backlogs have put huge pressure on aircraft maintenance, repair, and overhaul (MRO) operations. The use of generative AI in MRO will increasingly go hand-in-hand with predictive maintenance.

Gen AI possibilities in MRO include virtual AI maintenance-and-repair copilots to speed up the research and troubleshooting process, augmented-reality engineering tools, maintenance-report writers, and tools to help supply chain analysts spot potential supply problems before they become operational or safety risks.

AI in air traffic management 

Air traffic control towers are about 35% understaffed in the United States, and the controller shortage extends to Europe and beyond. While nobody’s suggesting that AI fully take over air traffic management, AI can do a lot to prevent congestion and delays and boost safety.

For a one-stop shop of the sorts of research under way in AI for air traffic management, look no further than Single European Sky ATM Research (SESAR).

SESAR is pursuing more than than a dozen AI programs for air traffic management, many of them with passenger safety implications. They include:

• taxiway inspection and runway monitoring;
• prediction and management of air traffic hotspots;
• smart allocation of airspace among controllers;
• decision support for safety-critical operations;
• improved adverse-weather forecasting;
• speech recognition to reduce manual inputs by controllers;
• and automated conflict detection and collision prevention.

Some of these programs have already borne fruit; with others, there’s much work to be done. Making anything fail-safe is a lofty goal, but AI has already surprised us. No doubt flying will remain vastly safer than drives to the airport.