FAA Launches AI Flight-Planning Tool as Controller Training Questions Persist

The Federal Aviation Administration has begun using an AI-supported flight-planning system around Washington, D.C., while questions remain about whether controllers were sufficiently involved or trained before the rollout. The FAA’s September 21 announcement says the system makes recommendations only: FAA personnel review its output, and local air-traffic leaders retain authority to accept or reject its scheduling suggestions.

That division of responsibility is central to understanding the Strategic Management of Airspace, Routes and Trajectories platform, or SMART. It is not controlling aircraft or replacing controllers. Instead, it combines approximately 200 data streams including weather, flight paths, traffic flow and controller staffing metrics to identify potential congestion and propose responses before disruptions develop.

Representative Don Beyer requested that the rollout be suspended, saying he had been informed that controllers were not consulted during SMART’s design and development or trained before adoption at local airports. That training claim has not been independently established. Separately, the National Air Traffic Controllers Association said it was not involved in the system’s design, testing or implementation and therefore could not assess its workforce effects at this stage.

Human authority does not eliminate implementation risk

Keeping people in command establishes an important operating boundary, but it does not by itself answer whether a decision-support system is ready for broader use. Human reviewers need to understand what information a recommendation incorporates, where its limits lie and how it should be handled when local conditions or professional judgment point toward a different choice.

This distinction matters because automation can influence operations without directly controlling equipment. A recommendation presented through an established FAA process may still affect which scheduling option receives attention, how quickly staff respond and how competing constraints are weighed. The relevant preparation is therefore not training someone to “fly” an AI system. It is ensuring that personnel can interpret, challenge and, when appropriate, reject its output without losing situational awareness or slowing time-sensitive coordination.

The limited Washington-area introduction gives the FAA a narrower environment in which to examine that human-machine relationship before national expansion. A staged deployment can reduce exposure to unknowns, but its value depends on measurable evaluation. Useful questions include whether recommendations are operationally feasible, whether they reach decision-makers with enough time to act, how often local leaders accept or decline them and whether the system improves disruption recovery without transferring workload elsewhere.

Fewer delays remain a goal, not a demonstrated result

The FAA says SMART can plan hours, days or weeks ahead by comparing expected demand with weather and airspace constraints. Earlier warning could give specialists more time to adjust schedules or routes rather than reacting after congestion has formed. For passengers, the intended result is fewer delays, cancellations, diversions and extended waits.

Those benefits have not yet been demonstrated at national scale. United Airlines CEO Scott Kirby offered notably conditional support, saying, “If it works as we think it can,” SMART could reduce delays and cancellations more than any change in decades. The CEOs of Delta, American and Southwest also publicly supported the initiative, while emphasizing collaboration or thoughtful implementation.

Claims that greater efficiency could lower travel prices require additional caution. Reduced fuel use, overtime or disruption costs could lower airline operating expenses, but the available information does not establish how much would be saved or whether savings would reach passengers. SMART also cannot create unlimited airport capacity, remove bad weather or resolve controller staffing shortages. It can help planners allocate available capacity, not erase the constraints producing congestion.

Scheduling recommendations also raise fairness questions

Airspace optimization is not purely a mathematical exercise because a systemwide improvement can distribute costs unevenly. Moving or rerouting flights may benefit overall traffic flow while imposing delay, fuel or scheduling consequences on particular airlines, airports and passengers. Local leadership’s ability to reject recommendations is therefore important, but nationwide deployment will also require consistent rules for deciding when an option is operationally acceptable and how competing users are treated.

The FAA’s stated boundary is clear: SMART cannot take control of an aircraft, and accountable human officials make the decisions. What remains unresolved is whether the people reviewing its recommendations received adequate preparation and whether the limited rollout will produce documented evidence of accuracy, workload effects and equitable operational results. Those are the tests that will determine whether SMART becomes a useful national planning layer rather than simply another stream of information in an already complex air-traffic system.

By Thomas Caldwell — AMI’s senior editor for mechanical and mobility engineering, covering vehicle electronics, systems integration, electrification, chassis systems, propulsion, and safety policy.