AI won’t be piloting American aircraft just yet, but it could soon be deciding when they take off or cross paths. The FAA is preparing a radical overhaul of the country’s air traffic management. And if AI took control of the skies? That’s what the Federal Aviation Administration (FAA) is working on. They are developing a futuristic program called SMART (Strategic Management of Airspace Routing Trajectories) that could revolutionize air traffic management in the United States.
Officially, this AI is not meant to replace air traffic controllers. But for the first time, AI systems will anticipate congestion, adjust aircraft trajectories, and optimize flows before takeoff. Behind this project are three tech giants: Palantir Technologies, Thales, and Air Space Intelligence.
For years, the US has been dealing with a chronic shortage of air traffic controllers, aging infrastructure, and a surge in air traffic. The political shock from a deadly collision between a commercial aircraft and a helicopter near Ronald Reagan Washington National Airport exposed the flaws in a system once considered one of the safest in the world.
According to Politico, US Transportation Secretary Sean Duffy is leading a massive multi-billion-dollar technological modernization plan supported by the Trump administration and the air traffic controllers’ union NATCA. The head of the FAA, Bryan Bedford, likens the American airspace to “a perpetual traffic jam in Los Angeles,” plagued by delays and cancellations every morning.
Facing concerns about automating the skies, the companies involved reassure that SMART is not meant to separate aircraft or manage critical safety functions. Todd Donovan, Thales VP for aerial mobility solutions in the Americas, emphasizes this.
The role of controllers would theoretically remain unchanged, with AI intervening upstream in strategic flow management. Algorithms would analyze vast amounts of data to recommend invisible adjustments to passengers to avoid congestion before it forms. This proactive approach aligns with predictive financial or logistics systems—better to prevent than react.
However, the integration of SMART into the FAA’s outdated IT infrastructure and its funding remain uncertain. The experimental phase is nearing completion, with a live demonstration potentially starting in September, followed by validation in 2026. But the specifics of SMART’s integration and budgeting are still unclear. Todd Donovan mentions that the project currently lacks an official budget line.
