The United States Federal Aviation Administration is preparing to deploy an AI system called SMART in the airspace above Washington, DC, before expanding it across all 29 million square miles of US national airspace. The rollout is not just a technology story. It is a window into how a critical public infrastructure agency is attempting to use artificial intelligence to compensate for a workforce it has spent decades failing to rebuild.
What SMART Actually Does, and What It Does Not
The SMART system is designed to predict air traffic flows and flag potential conflicts by analyzing a combination of airline schedules, weather data, airport capacity, and airspace conditions. Its stated goals include reducing fuel consumption, improving on-time performance, and accelerating recovery from weather disruptions and congestion events.
What SMART does not do is equally important to understand. Following weeks of confusion among airline industry stakeholders about the agency’s intentions, the FAA clarified that the system would not alter any existing procedures for air traffic controllers or airlines. It produces what the agency describes as “alternative route information,” delivered through systems already in use. Controllers remain in charge. The AI advises; humans decide.
This distinction matters because it defines the risk profile. Philip Mann, a principal consultant at Vector Strategic Consulting LLC with 17 years of prior FAA experience, has noted that the system’s risk is not any single prediction it might get wrong. The concern is that SMART is a national-scale system with AI components that carry more unknowns than anything the FAA has previously deployed. Limiting the initial launch to the Washington, DC, area, specifically to the three major airports there, reduces the surface area of those unknowns before the system faces the full complexity of national airspace.
Mann has also flagged an open question worth watching: whether the rollout remains limited to aircraft flying at 24,000 feet and above, as the FAA described in June. At that altitude, the system would be handling cruise traffic plus the climbs and descents feeding the metro airports, a meaningful but bounded scope.
A $875 Million Contract and a Workforce in Decline
SMART is part of a broader $875 million, 12-year contract awarded to Air Space Intelligence, a Boston-based company, in June. The contract also covers a Flow Management Data and Services system intended to replace the current infrastructure at the FAA’s Air Traffic Control System Command Center in Virginia. Mann has described that new system as the “backbone,” with SMART functioning as the “predictive layer above it.”
Air Space Intelligence already operates a separate platform called Flyways, which the company claims helps manage over 40 percent of all US air traffic through partnerships with airlines including Alaska Airlines. That platform uses a four-dimensional digital twin of US national airspace to model traffic and weather conditions.
The timing of this deployment is not incidental. The FAA is managing a structural staffing crisis that traces back to 1981, when President Ronald Reagan dismissed more than 11,000 air traffic controllers. The workforce has never fully recovered. According to a US Government Accountability Office report published in December 2025, the overall air traffic controller workforce declined by 6 percent over the past decade. Hiring efforts were disrupted by government shutdowns in 2013 and 2018 to 2019, and by training restrictions during the COVID-19 pandemic.
The situation worsened further in 2025. Early in the second Trump administration, hundreds of probationary FAA hires were laid off as part of the cost-cutting effort led by Elon Musk’s Department of Government Efficiency. Air traffic controllers themselves were spared, but some employees working on air traffic control systems were affected. A federal judge later ordered the reinstatement of 132 of those employees. A 43-day government shutdown that ran from October 1 to November 12, 2025, added further strain: Transportation Secretary Sean Duffy stated that 15 to 20 controllers were retiring each day during the shutdown, compared to an average of four per day. FAA chief Bryan Bedford described the agency as losing up to 500 air traffic controller trainees during that period.
In May of this year, the FAA reduced its estimated staffing requirements for the 2026 to 2028 period by approximately 2,000 controllers. That reduction is a signal worth reading carefully.
Why This Moment Is Different From Other AI Deployments
Most AI deployments carry consequences that are recoverable. A wrong recommendation in a retail system costs a sale. A flawed prediction in a logistics platform delays a shipment. Air traffic management does not offer that margin. The January 29, 2025, collision between a US Army Black Hawk helicopter and an American Airlines passenger jet near Ronald Reagan Washington National Airport killed 67 people. Investigators identified airport tower understaffing as one of multiple causal factors, with a single controller managing both helicopter and runway traffic that night.
That context reframes what SMART represents. This is not an agency experimenting with AI because the technology is fashionable. It is an agency under pressure, deploying AI tools while simultaneously replacing radar systems dating to the 1980s, radios, telecommunication lines, and the voice switches connecting controllers to pilots. The modernization effort is multi-billion-dollar in scale and decades overdue.
Mann has stated he will be watching for what evidence gates the next expansion: performance under real operational workload and degraded data conditions, not demonstration environments. He has also raised a governance question that applies to any high-stakes AI deployment: who owns the outcome when a prediction is wrong?
That question does not yet have a written answer.
In Short
SMART is an AI advisory system, not an autonomous decision-maker. It predicts, suggests, and informs. Controllers retain authority. The Washington, DC, launch is a deliberate scope limitation designed to reduce unknowns before a national rollout. Behind the technology story is a workforce crisis decades in the making, a modernization effort long overdue, and a set of accountability questions that the FAA has not yet answered publicly. The stakes are not abstract: the consequences of getting this wrong are measured in lives.
Based on reporting from Ars Technica.