The Federal Aviation Administration is preparing to begin initial operations of an AI-enabled air-traffic planning system in Washington-area airspace, a limited first deployment that is expected to precede a broader national rollout. The system, called SMART—short for Strategic Management of Airspace, Routes, and Trajectories—is intended to advise on traffic-flow planning before aircraft depart, rather than automate the real-time work of directing flights.
Reporting has identified the first deployment area as the three principal airports serving the capital region: Ronald Reagan Washington National, Washington Dulles International and Baltimore/Washington International. But the FAA’s public materials describe the timing more cautiously: its SMART fact sheet says initial operations begin in fall 2026. As of September 21, the agency had not published a separate newsroom announcement confirming that a reported Washington trial had gone live.
Planning support, not autonomous control
The distinction is important. SMART is not presented as an autonomous controller. The FAA says air traffic controllers remain responsible for keeping aircraft safely separated, while SMART operates as a planning layer alongside existing agency systems. Its intended contribution is earlier visibility into conditions that can create delays or congestion, giving controllers, airlines and other operators more time to consider routes, departure times and other adjustments.
According to the FAA’s fact sheet, the cloud-based platform uses AI to analyze airline schedules, weather, airport capacity, airspace conditions and operational constraints. It is designed to predict traffic flows, identify potential conflicts and continuously update forecasts and planning recommendations as conditions change. The agency’s stated premise is that operational conflicts are often recognized only when flights are already at the gate, taxiing or airborne—when the available options are narrower and the effects can spread through the network.
Part of the FMDS modernization effort
SMART is an enhancement within a larger Flow Management Data and Services, or FMDS, program. FMDS is meant to become the technological backbone for the FAA’s Air Traffic Control System Command Center, balancing demand with capacity and providing the data used in traffic management. In June, the FAA announced that it had awarded Air Space Intelligence a contract covering the two technologies. Independent reporting has put the award at $875 million over 12 years.
The intended division of labor matters for both aviation and enterprise AI. SMART’s focus is strategic traffic-flow management: looking across schedules, capacity constraints and forecast conditions to help address bottlenecks before they cascade. That is different from handing an AI system control over separation decisions in the moment. The FAA says the software will help create a shared data-driven view for operators and controllers, but responsibility for operational decisions and safe separation remains with the human controller.
Human factors and operational safeguards
That design reflects a practical constraint of deploying AI in a safety-critical environment. A prediction can be useful only if a human operator can understand when to rely on it, how it fits with other information and what to do when circumstances change. The FAA’s own human-factors work identifies decision-support tools as a central use case for AI and machine learning in air traffic control and traffic-flow management. The agency also notes that adoption depends on whether human operators can use new capabilities effectively, with clear roles, appropriate reliance and resilience when technology fails or produces an unhelpful recommendation.
For SMART, those safeguards will be as consequential as forecast quality. The public materials describe what the system is intended to ingest and produce, but do not spell out public evaluation thresholds for the Washington operations. A limited deployment can therefore serve a dual purpose: testing whether the forecasts help planning and testing whether the information is presented in a way that fits controller and airline workflows without adding confusion or encouraging over-reliance.
Path toward a wider rollout
Ars Technica reported that the Washington launch was expected to be the first step toward a nationwide deployment across the U.S. national airspace system. The FAA has said its new platforms are intended to bring critical data into one environment, including visualization of weather patterns and flight paths, so the agency can identify delays and unused capacity earlier. The broader FMDS replacement effort follows the FAA’s 2025 vendor challenge, which described the existing Traffic Flow Management System as decades old and facing performance and maintainability problems.
A wider rollout would make SMART a notable example of government AI moving beyond pilots and back-office analysis into operational decision support. Its success, however, will not be measured merely by whether it can generate forecasts. In this setting, the harder standard is whether its recommendations improve planning while preserving clear human accountability, effective coordination among airspace users and safe recovery when predictions do not match reality.
The Washington deployment is consequently more than a technology launch. It is an early operational test of how an AI planning layer can be inserted into an already complex human system—one where the agency’s aim is to reduce congestion before flights depart, while ensuring that the people responsible for safety remain in control.
