Lawmakers Seek GAO Audit of FAA's AI Air Traffic Software
House Democrats cite safety and controller staffing concerns after the agency piloted two AI-enabled traffic platforms
By AXY Government Desk ·
Key points
- More than 20 members of Congress asked the GAO to audit the FAA's AI-enabled air traffic software, FMDS and SMART.
- Lawmakers earlier asked the FAA and the Transportation Department to pause the systems above the National Capital Region until more safety measures are in place.
- Lawmakers say there has been little transparency about how the systems were tested and why the NCR was chosen as the pilot site.
- The debate combines AI safety, cybersecurity findings from the GAO and a shortage of air traffic controllers.
House Democrats are asking the Government Accountability Office to audit the Federal Aviation Administration's use of artificial intelligence in air traffic software, citing safety concerns and a shortage of air traffic controllers. According to Federal News Network, more than 20 members of Congress signed a letter asking the GAO to review two FAA systems, the Flow Management Data and Services (FMDS) and Strategic Management of Airspace, Routes and Trajectories (SMART). The request came only a month after the agency piloted the platforms, and it is the second letter the agency has received from lawmakers in the past month raising concerns about them.
The FAA is the federal agency responsible for civil aviation safety and for the national air traffic control system that guides commercial, cargo and general aviation flights across US airspace. Its use of artificial intelligence is now drawing scrutiny from two directions: the agency's own inspector general and lawmakers on Capitol Hill. An inspector general is an internal watchdog within a federal agency that reviews its programs and operations, while the GAO is the nonpartisan audit arm of Congress that examines how federal money is spent and how government programs perform. A GAO audit would give lawmakers an independent assessment of how the systems were tested, approved and deployed.
What the lawmakers are asking for
The letter to the GAO focuses on FMDS and SMART, which the source material describes as AI-enabled air traffic software. By their names, both systems relate to managing traffic flow and the routing of aircraft through the airspace, although the source material does not describe in detail how each one operates. Lawmakers have not argued that the technology should be abandoned. Their stated position is that the agency's modernization of the air traffic system is necessary, but that safety must come first.
The push for a review follows an earlier letter sent in the same period. In that letter, members of Congress wrote to Transportation Secretary Sean Duffy and FAA Administrator Bryan Bedford asking the agency to pause use of FMDS and SMART in the airspace above the National Capital Region (NCR), the area around Washington, until the agency put more safety measures in place. The NCR is one of the most congested and sensitive airspace zones in the country, which is one reason lawmakers are watching closely how new systems are introduced there.
Why the earlier letter about the National Capital Region matters
In a letter dated Oct. 9, lawmakers wrote: “There has been little transparency from either DOT or the FAA on how the AI systems underlying SMART were tested … nor has information been shared as to why the NCR was chosen as the pilot test region.” The complaint goes to the heart of the oversight question. Lawmakers are not only asking whether the systems work, but whether the public and Congress have been given enough information to judge whether they are safe to operate in a busy region.
The Department of Transportation oversees the FAA, so the questions raised about testing and site selection fall to both agencies. Federal News Network reported that the lawmakers' concerns were expressed in both recent letters, which argued that without adequate review, AI could affect the current control systems in ways that are not yet understood. For a pilot program, the choice of test location can shape the risk profile of an experiment, since a system tested in a dense, high-stakes airspace carries different consequences than one tested in a quieter area.
Cybersecurity and the GAO's earlier findings
Lawmakers also raised cybersecurity concerns, pointing to recent reports of AI models behaving unpredictably. The source material does not detail those reports, so their scope cannot be assessed here. But the cybersecurity issue has a separate and documented background. The GAO released two reports on aviation cybersecurity, one in June and one in August, which found that the FAA's air traffic and data communication systems are vulnerable to cyber threats. The GAO recommended improvements to threat detection and mitigation efforts.
Jennifer Franks, director of the GAO Center for Enhanced Cybersecurity, told Federal News Network in August that the agency's recommendations were aimed at closing gaps in how controls are put into practice. She said: “You can have a procedure and guardrails in place, but unless you fully implement all the necessary controls, you will have gaps in your processing of capabilities.” She also pointed to tighter management of privileged user access and to zero trust architecture, a security approach that assumes no user or device should be trusted automatically and requires verification for each access request.
For a non-specialist reader, the connection is straightforward. Air traffic systems are networked, and any new software that influences how flights are managed adds another layer that must be protected. A new AI tool that is not fully understood, or that is introduced without complete controls, could widen the points an attacker or a fault could exploit. That is why lawmakers are linking the AI questions to the cybersecurity findings already on the record.
The workforce question and controller concerns
The debate also cannot be separated from the FAA's workforce. The source material notes that the controversy is unfolding alongside a shortage of air traffic controllers, a staffing problem that has been widely reported in recent years. Air traffic controllers are the people who direct aircraft, maintain separation between them and respond to emergencies, and the system depends on having enough trained staff on duty.
The National Air Traffic Controllers Association (NATCA), the union representing controllers, told Federal News Network that any new technology should “complement and not replace the experience, training and professional judgment of air traffic controllers who are responsible for the safety of the National Airspace System.” That position reflects a broader tension in aviation modernization. Technology can help manage workload and traffic, but if it is seen as a substitute for experienced staff, it can raise questions about whether controllers will be asked to rely on tools they cannot fully check.
For passengers, the stakes are practical. Delays, route changes and the speed at which controllers can respond to unexpected events all depend on how the system is staffed and equipped. Shortages can strain controllers, and new technology that changes how they work must be introduced carefully so that it supports rather than complicates their jobs.
Automation bias and the case for caution
The lawmakers' letters draw on a concept that has been studied in aviation and other fields. The letter states: “Decades of research show that people tend to over-rely on automated recommendations, a phenomenon known as automation bias, particularly under overworked conditions.” Automation bias describes the tendency of people to accept a machine's suggestion even when their own judgment or other evidence points elsewhere. The risk is greater when people are busy or fatigued, which is why the staffing shortage and the introduction of new tools are being considered together.
The letter adds that adding new, untested layers to these systems must be done in partnership with all stakeholders in the aviation safety ecosystem. It also cites airline industry officials who have said the technology is not yet ready to immediately transform flight operations. That statement, as quoted in the letter, suggests that industry groups share some of the caution expressed by lawmakers, though the source material does not provide the full context of those officials' remarks.
The approach lawmakers are urging is incremental: test carefully, disclose how systems were evaluated, involve controllers and airlines, and keep human judgment central. Whether the FAA's current pilot programs meet that standard is the question the requested GAO review would be designed to examine.
What happens next
The GAO has not, in the material provided, announced whether it will accept the request or when any review would begin, so the timing and scope of an audit remain unknown. An audit would typically examine testing, governance and risk controls, and could make recommendations that the FAA would be expected to respond to. Separately, the FAA's inspector general is already examining the agency's use of AI, which could produce its own findings.
Several developments are worth watching. The first is whether the FAA responds to the request to pause FMDS and SMART in the National Capital Region. The second is whether the Department of Transportation and the FAA release more information about how the systems were tested and why the NCR was selected. The third is whether the GAO's earlier cybersecurity recommendations are implemented alongside any new AI deployments. Taken together, these threads will determine whether the debate over AI in air traffic control moves from letters and reports toward concrete changes in how the system is tested and run.
Topics: Federal Aviation Administration, Government Accountability Office, Air traffic control, Artificial intelligence, Air traffic controllers, Aviation cybersecurity, Congress, National Capital Region
Sources: Federal News Network
https://axynewsglobal.com/article/lawmakers-seek-gao-audit-of-faa-s-ai-air-traffic-software