AI is entering air traffic control. Is it taking over the tower?
AI may soon help decide when flights move—but the person separating aircraft is still human.
GOT 30 SECONDS?
The short answer
No. The new system is designed to analyse schedules, weather, airport capacity and airspace conditions, then help traffic managers spot congestion before aircraft depart. It can recommend a better plan. Human controllers still issue clearances and remain responsible for keeping aircraft safely separated.
So what is actually changing?
The United States Federal Aviation Administration has awarded Air Space Intelligence a 12-year contract worth up to US$875 million for two connected programmes: FMDS and SMART. The ambition is to replace older traffic-flow tools with one platform that can see the national picture earlier.
Think of it as moving from reacting to a traffic jam after it forms to spotting where the jam is likely to form—and adjusting the plan before dozens of aircraft join it.
What will the AI look at?
The system is expected to combine airline schedules, weather, airport capacity, airspace conditions and operating constraints. It can model how one disruption may spread across routes and airports, then identify possible conflicts or bottlenecks.
That matters because a thunderstorm over one busy corridor can affect flights far beyond the storm itself. A better national view could help airlines and traffic managers choose smarter departure times or routes before the disruption multiplies.
Will AI talk directly to pilots?
That is not what has been announced. SMART is a traffic-flow planning and decision-support system—not an autonomous controller replacing the person on the radio.
Controllers still manage aircraft in real time, issue clearances and apply separation rules. AI may improve the information and options in front of them, but advice is not authority.
Could it reduce delays?
Possibly. Earlier warnings and better coordination could prevent some avoidable congestion, especially when weather, demand and limited airport capacity collide. The FAA has said the tools are intended to improve traffic flow, reduce congestion and strengthen resilience.
But software cannot create a runway, remove a thunderstorm or instantly solve staffing shortages. A recommendation is useful only when the data is reliable, the plan is practical and the people running the system understand why it makes sense.
The real question is who can challenge the machine
In safety-critical work, “the computer said so” is never enough. Operators need to know what information the system used, what it may have missed and when a human should reject its recommendation.
The strongest future is not AI versus controllers. It is AI finding patterns at a scale no person can scan alone, while trained humans keep authority, context and accountability.
TAKE THIS WITH YOU
AI is not taking over the tower. It is becoming a powerful planning partner—and humans still need the authority and confidence to question it.
WATCH · MIRA EXPLAINS
AI is entering the tower.
Is it replacing the controller—or helping humans see trouble sooner?
Read the plain answer ↑Sources
Used to check the facts in this story.