FAA Debuts SMART: How AI Is Changing U.S. Air Traffic Management

What SMART Actually Does

SMART stands for Strategic Management of Airspace Routing Trajectories. The FAA has deployed it across three Washington-area airports: Ronald Reagan Washington National, Dulles International, and Baltimore/Washington International.

The system pulls in real-time flight data and applies predictive analytics to help controllers make faster, better-informed decisions. Two specific use cases stand out:

• Weather-related ground delays: When weather moves faster than controllers can respond, SMART can flag aircraft ready to move before a controller has bandwidth to act.
• Overlapping arrival windows: When multiple flights are converging on the same airport at the same time, SMART can recommend subtle route adjustments to space out landings.

FAA Administrator Bryan Bedford described it as giving controllers “much better decision-making capabilities” — not replacing their judgment, but sharpening it.

Human Controllers Stay in Charge

This is worth stating clearly: SMART is a decision-support tool, not an autonomous system. Transportation Secretary Sean Duffy was direct about it at the launch press conference.

“It’s always going to be a human that manages the airspace in America,” Duffy said.

That framing matters. The aviation industry has been cautious about AI adoption, and some airline officials were reportedly hesitant about the program before a series of meetings with FAA leadership helped ease concerns. The “human in the loop” positioning appears to be a deliberate choice to build trust with both operators and the public.

The Rollout Timeline

SMART isn’t going nationwide overnight. Here’s how the phased approach looks:

• Now: Live at the three Washington-area airports
• Next 90 days: Testing period to evaluate real-world performance
• After testing: Expansion to additional U.S. locations
• By next summer: Passengers should start noticing improvements, according to Duffy

That’s a measured pace — which is appropriate for safety-critical infrastructure. The 90-day testing window gives the FAA real operational data before scaling.

The Infrastructure Problem SMART Can’t Fix

Here’s the tension that Monday’s launch made impossible to ignore. While Bedford was showcasing SMART at a DOT press conference, a primary circuit failure at a Philadelphia air traffic control facility knocked out operations across Newark, JFK, and LaGuardia. A backup fiber line was found severed — reportedly cut during Amtrak construction between New Brunswick and Newark, New Jersey.

Bedford called the fiber break “massive” and estimated a 13-hour repair window.

SMART is designed to optimize how controllers work within the existing system. It can’t patch aging hardware or prevent a construction crew from cutting a critical cable. The FAA is running a modern AI layer on top of infrastructure that, in some cases, is decades old. That gap is the bigger story.

Why This Matters for AI Watchers

SMART is a useful case study in how AI gets deployed in high-stakes, regulated environments. A few things worth noting:

Adoption required relationship-building. Bedford met with airline CEOs on September 10 specifically to demonstrate the system. That kind of stakeholder management is often what determines whether an AI tool actually gets used — not just whether it works technically.

The “human in the loop” framing is doing real work. In aviation, public trust is non-negotiable. Positioning AI as a tool that enhances human controllers rather than replacing them isn’t just PR — it’s the architecture of responsible deployment in safety-critical contexts.

Real-world testing before scaling is the right call. A 90-day live pilot at three airports before broader rollout is exactly how consequential AI systems should be evaluated. Benchmark demos are not the same as operational performance.

The Practical Takeaway

SMART represents a real step forward for U.S. air traffic management — but it’s one piece of a much larger puzzle. If the testing period produces measurable reductions in departure and arrival delays at Washington-area airports, the case for broader rollout becomes much stronger.

For anyone tracking AI adoption in regulated industries, this is a useful model to watch: phased deployment, human oversight built in from the start, and a clear stakeholder communication strategy. Whether the results match the promise is something we’ll know more about in 90 days.