AI can be predictive. It can spot a conflict before it becomes one.
In an era when the skies grow ever more crowded and the margin for human error ever more consequential, the Federal Aviation Administration is turning to artificial intelligence to serve as an unwavering second set of eyes over the nation's busiest runways. Three firms — ASI, Palantir, and Thales — are competing to build the SMART Airspace Tool, a system designed to detect dangerous aircraft spacing before controllers must react under pressure. The move reflects a quiet but profound admission: that the complexity of modern airspace may be outpacing the limits of unaided human attention, and that the next chapter of aviation safety may be written not by faster reflexes, but by smarter foresight.
- Runway close calls at major airports have accumulated into a pattern the FAA can no longer attribute to isolated human error — the system itself is straining under the weight of modern air traffic volume.
- Three technology firms are locked in competition for a contract that could define how artificial intelligence enters one of the most safety-critical environments in public infrastructure.
- September demonstrations will force a concrete reckoning: which system can translate raw flight data into actionable alerts fast enough to matter when aircraft are seconds apart on final approach.
- The FAA is betting that AI augmentation — layered atop aging radar and human expertise — offers a faster and more practical path to modernization than replacing legacy infrastructure wholesale.
- If the SMART Airspace Tool succeeds, it may set a global precedent, accelerating AI adoption by aviation authorities worldwide and redefining the standard of care for air traffic management.
The Federal Aviation Administration is nearing a decision that could fundamentally change how air traffic controllers manage aviation's most dangerous moments — the approach and landing phases, when aircraft are closest together and the cost of error is highest. The agency is evaluating AI systems designed to catch problems before they become incidents, targeting the close calls that, left unaddressed, can accumulate into the conditions that precede accidents.
Three companies are competing for the contract: ASI, currently considered the frontrunner, along with Palantir and Thales. The prize is the SMART Airspace Tool — a system that would process real-time radar data, flight plans, and aircraft positions to alert controllers to potential conflicts before situations become critical. It is conceived not as a replacement for human judgment, but as a tireless complement to it, one immune to fatigue and capable of tracking more variables simultaneously than any individual controller.
Runway incidents have long anchored safety investigations. These are the moments of smallest separation and largest consequence — situations where aircraft came closer than regulations permit, or where emergency instructions were the only barrier to collision. Most resolve without injury, but each represents a failure point: a moment when something nearly went wrong and the system only barely held.
Demonstrations of the competing systems are expected in September, a timeline reflecting the FAA's urgency. The agency has faced years of pressure to modernize infrastructure that, while reliable, is aging. AI offers a pragmatic middle path — augmenting existing systems rather than replacing them, catching what the current setup might miss.
The broader stakes extend well beyond any single airport. Aviation's extraordinary safety record is built on preventing incidents before they occur, and every undetected close call is a missed signal. If the SMART Airspace Tool performs in demonstration, it could become a template for air traffic control modernization both domestically and abroad — a proof of concept that the next frontier of aviation safety lies not in faster human reactions, but in systems that see the danger coming first.
The Federal Aviation Administration is moving toward a decision that could reshape how air traffic controllers manage the most dangerous moments in aviation: the approach and landing phases when aircraft are closest together and mistakes carry the highest cost. The agency is evaluating artificial intelligence systems designed to catch problems before they become incidents—near-misses that, left unaddressed, can accumulate into patterns that precede accidents.
Three companies are in the running for what the FAA calls the SMART Airspace Tool, a system meant to process real-time flight data and alert controllers to potential conflicts or unsafe spacing between aircraft. ASI appears to be the frontrunner, though Palantir and Thales remain in contention. The competition reflects a broader recognition within aviation that the current system, built on decades of human expertise and radar technology, has limits. Controllers manage thousands of flights daily across the nation's busiest airports. They are skilled, but they are also human—subject to fatigue, distraction, and the simple fact that more aircraft in the sky means more variables to track simultaneously.
Runway incidents have long been a focus of safety investigations. These are the moments when aircraft are on final approach or crossing active runways, when separation margins are smallest and the consequences of error are largest. The FAA has documented close calls—situations where aircraft came closer than regulations allow, or where controllers had to issue emergency instructions to prevent collision. Most are resolved without injury, but each one represents a failure point in the system, a moment when something nearly went wrong.
The SMART Airspace Tool is designed to function as a second set of eyes, one that never blinks and never gets tired. It would analyze radar data, flight plans, and aircraft positions to identify potential conflicts and alert controllers before the situation becomes critical. The system would flag spacing violations, runway incursions, and other hazards in real time, giving controllers the information they need to make corrections while they still have options.
Demonstrations of the competing systems are expected in September, a timeline that suggests the FAA wants to move quickly. The agency has faced pressure to modernize its air traffic infrastructure for years. The current system relies on technology that, while reliable, is aging. Modernization efforts have been slow and expensive. AI offers a different approach: rather than replacing the entire system, it can augment it, working alongside existing infrastructure and human controllers to catch problems the current setup might miss.
The stakes are significant. Aviation is statistically one of the safest forms of transportation, but that safety record is built on systems designed to prevent incidents before they happen. Every close call that goes undetected is a missed opportunity to identify a problem. Every pattern that goes unrecognized is a risk that compounds. The FAA's turn toward AI reflects a judgment that the agency has reached the limits of what human attention alone can achieve in an increasingly crowded airspace.
If the SMART Airspace Tool proves effective in demonstration, it could become a model for other airports and other aspects of air traffic control. The success or failure of this effort will likely influence how quickly other aviation authorities around the world adopt similar systems. For now, the three competing firms are preparing their pitches, and the FAA is preparing to see which system can best translate data into the kind of actionable intelligence that keeps aircraft safely separated in the sky.
A Conversa do Hearth Outra perspectiva sobre a história
Why does the FAA need AI to catch runway problems? Don't controllers already have systems in place?
They do, but those systems were built for a different era. A controller managing approach traffic at a major airport is tracking dozens of aircraft simultaneously, each with its own speed, altitude, and clearance. The radar and radio systems work, but they're reactive—a controller sees a problem and responds. AI can be predictive. It can spot a potential conflict before it becomes one.
What kinds of incidents are we talking about? How often do these close calls actually happen?
The FAA doesn't release exact numbers, but runway incursions and spacing violations happen regularly at major airports. Most are caught and corrected without incident, but that's the point—they're near-misses. Each one is a moment where the system worked because someone was paying attention. AI is meant to catch the ones that slip through.
Three companies competing—why does it matter which one wins?
Because the system that gets chosen will likely become the standard. It shapes how air traffic control evolves, what data gets prioritized, how alerts are designed. The winning system will influence safety practices across the country.
What happens if the AI makes a mistake? What if it flags something that isn't actually a problem?
That's the real challenge. Too many false alarms and controllers stop trusting the system. Too few and you miss actual hazards. The demonstrations in September will test whether these systems can find the right balance—catching real problems without crying wolf.
Is this the future of air traffic control?
It's the beginning of it. Controllers won't disappear. But the job will change. Instead of watching radar and listening to radio, they'll be managing a system that's constantly analyzing and alerting. The human judgment stays central. The AI just gives them better information to judge with.