DOT pilots SMART AI tool to cut US flight delays

The US Department of Transportation is moving ahead with a pilot of a new artificial intelligence supported air traffic management platform intended to spot bottlenecks earlier in the day and reduce the flight delays and cancellations that have become a persistent frustration for travelers across the United States.

New AI platform built around congestion forecasting

Publicly available information from the Department of Transportation and Federal Aviation Administration describes the new tool, known as Strategic Management of Airspace, Routes and Trajectories, or SMART, as a next generation planning platform that concentrates on when and where congestion is likely to emerge rather than only reacting once delays have already built up.

SMART is described as ingesting roughly two hundred streams of operational data, including airline schedules, weather forecasts and live weather radar, filed and active flight plans, traffic flow information, airport capacity constraints, airspace restrictions, and staffing related metrics at key facilities. By combining those inputs, the system generates a common picture of expected demand on the air traffic system over the course of the day.

The system’s AI supported analytics are intended to highlight choke points in advance, such as mid afternoon periods when planned flight volumes exceed the practical throughput of a major hub during forecast thunderstorms, or when en route sectors are expected to reach saturation as traffic is funneled around convective weather or military airspace.

Instead of waiting for those strains to appear in real time, traffic managers, airline operations centers, and airport stakeholders are expected to use SMART’s projections to make earlier decisions about reroutes, schedule adjustments, or ground delay programs that are more targeted and less disruptive.

Pilot rollout centered on the Washington, D.C. region

Reports indicate that the initial deployment of SMART is concentrated in the National Capital Region, which gives the Department of Transportation and FAA an opportunity to test the tool in a complex environment with three major commercial airports and dense traffic flows along the busy Northeast Corridor.

Washington Reagan National, Washington Dulles International, and Baltimore Washington International airports all sit within constrained airspace that can become quickly saturated during convective weather or traffic surges. SMART is being positioned as a way to better understand how schedules, airspace constraints, and weather will interact across this multi airport region before problems cascade throughout the day.

According to published coverage, the pilot phase will be used to validate the quality of SMART’s forecasts against actual conditions, refine how the tool interfaces with existing traffic management systems, and develop playbooks for how airlines and traffic managers respond to the system’s alerts. A staged national rollout is expected to follow if the pilot demonstrates consistent benefits without compromising safety or controller workload.

DOT materials characterize this as one element of a broader modernization effort that also includes runway surface awareness initiatives at more than one hundred airports and updated tower surface management tools that are being deployed in phases through the middle of the decade.

How SMART fits into the NextGen modernization push

The SMART pilot is being framed within the wider NextGen air traffic modernization program, which has focused for more than a decade on shifting the United States from a ground based radar system toward satellite enabled navigation, digital data exchange, and trajectory based operations that treat each flight as a four dimensional path defined by position and time.

Existing systems such as En Route Automation Modernization and Time Based Flow Management already provide controllers and planners with conflict detection and high level flow management functions, but they were originally built around more deterministic, rules based models. SMART is described as an overlay that uses advanced data fusion and AI supported pattern recognition to search across the entire system for combinations of factors that historically have led to extended delays.

In that sense, the new platform does not replace core separation and safety functions. Instead, it is intended to give decision makers better situational awareness about how today’s schedule, today’s weather, and today’s constraints are likely to interact, and which levers are available to reduce imbalance between demand and capacity before gridlock develops.

The Department of Transportation’s published artificial intelligence strategy outlines a broader interest in using machine learning to assist human operators in complex environments, including air traffic management, while emphasizing that final authority remains with certified personnel and that any safety critical functions must meet existing certification standards.

Potential impact for travelers and airlines

For travelers, the near term effect of the SMART pilot is expected to show up less as a visible new interface and more as changes in when and how airlines and the FAA implement flow control measures. If congestion can be anticipated earlier in the day, schedules could be adjusted before long lines of aircraft build up on taxiways, reducing the number of flights that sit for extended periods waiting for takeoff clearance.

Airlines may also be able to use SMART’s systemwide view to select more efficient reroutes or to proactively trim portions of their schedule on days when weather or airspace constraints sharply reduce capacity at critical hubs. That can still mean cancellations, but the intent is to concentrate disruptions earlier and avoid the rolling knock on effects that can persist for days after a severe disruption.

For airports, particularly those serving as connecting hubs, improved predictability can translate into better gate management, more orderly passenger flows through security and concessions areas, and more efficient coordination with ground handling and deicing providers on peak demand periods.

From a consumer protection standpoint, the Department of Transportation’s inspector general has previously pointed to limitations in how delay and cancellation data are collected and analyzed across different systems. The introduction of tools that more tightly integrate operational and planning data could help agencies, airlines, and airports understand which bottlenecks are most responsible for disruptions and where investments in infrastructure or procedures may yield the greatest benefit.

Questions and safeguards around AI in aviation

The deployment of AI supported tools into air traffic management is drawing close scrutiny from aviation labor groups, safety advocates, and researchers, who have stressed that any use of machine learning must be transparent, explainable, and subject to rigorous validation in a safety critical environment.

Recent submissions to federal dockets on AI in transportation highlight concerns that opaque algorithms could make it harder for controllers and pilots to understand why a particular recommendation is being made, potentially creating new kinds of operational risk if decision support tools are trusted without adequate human oversight.

Academic work published in 2026 on decision assurance layers for AI assisted flight planning points to the need for mechanisms that continuously check automated recommendations against established safety constraints and operational rules, with clear ways for human operators to override suggestions and provide feedback that improves the system over time.

DOT’s own strategy documents describe AI as a means of reducing manual workload and improving data driven decision making but emphasize that the agency is pursuing a cautious, iterative approach, starting with advisory functions and analytics rather than direct control of aircraft trajectories. The SMART pilot fits that pattern by focusing on strategic traffic planning and allowing air traffic controllers and airline dispatchers to remain the final arbiters of how flights are managed on any given day.