Transportation Secretary Sean Duffy is pushing to rebuild the nation’s air traffic control system from scratch, and his message to the roughly 14,000 controllers who keep American skies safe is blunt: artificial intelligence will help you do your job, not take it.
The Department of Transportation and the Federal Aviation Administration in April 2026 posted a formal Request for Solutions on SAM.gov, the federal contracting portal, seeking a single “Prime Integrator” to design and deliver an entirely new air traffic control platform. The move signals that what began as exploratory outreach to industry has become a live procurement, with companies now able to submit formal proposals under binding federal acquisition rules.
In public communications tied to the rollout, Duffy and FAA leadership have consistently described new technology, including AI, as a support layer for human operators rather than a path toward autonomous control. The DOT framed the initiative as a call for “top innovators” to overhaul infrastructure that still relies on decades-old software, while emphasizing that controllers remain central to the system’s safety architecture.
Why the overhaul is happening now
The current air traffic control network handles roughly 45,000 flights a day using technology that, in some facilities, predates the internet. Previous attempts to modernize, most notably the multibillion-dollar NextGen program launched in the mid-2000s, were plagued by shifting requirements, vendor coordination failures, and repeated schedule slips that drew sharp criticism from the Government Accountability Office.
Compounding the technology problem is a staffing crisis. The FAA has operated below its own controller workforce targets for years, with training pipelines struggling to keep pace with retirements. That shortage is a key reason the AI question matters so much to controllers and their union: any suggestion that automation could thin already-stretched ranks touches a raw nerve.
Funding for the new effort is tied to the legislative package known as the One Big Beautiful Bill, which includes provisions for transportation infrastructure investment. By selecting a single integrator rather than splitting work among multiple vendors, the administration is betting that centralized responsibility will deliver faster results and clearer accountability than the fragmented approach that hampered NextGen.
What AI might actually do in the new system
Neither the Request for Solutions nor the earlier Request for Information spells out exactly how AI algorithms would fit into controller workflows. The design details will likely emerge only after a contractor is chosen and begins engineering work alongside regulators, controllers, and safety specialists.
But the FAA’s recent track record offers clues. The agency has already been deploying surface safety technology at airports nationwide, giving controllers real-time alerts to help prevent runway incursions and ground-level hazards. Those tools use automated data fusion and warning systems, but every final call still belongs to a human. The pattern across recent FAA deployments is consistent: software monitors, flags, and recommends while people decide.
In a fully modernized system, AI could handle routine sequencing of arriving aircraft, scan continuously for potential conflicts and surface them for human review, optimize routes for fuel savings, or fuse data from radar, satellite, and ground sensors into a single coherent picture. It could also help integrate new entrants into the airspace, including commercial drones and urban air taxis, without overloading controllers who are already managing heavy traffic loads.
The single-integrator gamble
Choosing one contractor to own the entire build is a deliberate break from past practice. Centralizing the work could eliminate the finger-pointing and incompatible subsystems that derailed earlier multi-vendor programs. But it also concentrates enormous risk. If that contractor stumbles, the whole program could stall.
The model likely favors established defense and aerospace firms such as Raytheon, Leidos, or L3Harris, all of which already hold major FAA contracts and have the scale to manage a project this large. Smaller AI-focused companies, which often drive the fastest advances in machine learning, could find themselves limited to subcontractor roles with little influence over core architecture. Whether the final selection encourages partnerships that blend traditional aviation experience with cutting-edge AI research will say a lot about the administration’s real priorities.
What controllers and their union are watching
The National Air Traffic Controllers Association, which represents the controller workforce, has not released a public statement specific to this procurement. But the union has historically pushed back hard on any modernization proposal it views as a threat to staffing levels or working conditions. If NATCA perceives AI tools as a stepping stone toward reducing headcount, it will almost certainly press for contractual protections, mandatory training programs, and explicit limits on machine-led decision-making.
That dynamic could shape the project’s political trajectory. Congressional support, contract terms, and even the pace of deployment may hinge on whether organized labor sees the effort as genuinely controller-centered or as a long-term automation play dressed in reassuring language.
What it means for travelers
For the flying public, the stakes are practical. A successful rebuild could mean fewer weather-related delays, better safety margins during peak traffic, and a network resilient enough to absorb the growth airlines are banking on as passenger demand continues to climb. A botched rollout, on the other hand, could disrupt operations at the worst possible time.
What Duffy and the FAA have communicated so far points in one direction: technology should make controllers more effective, not expendable. The verified record shows a government willing to commit to large-scale change while publicly affirming that human expertise stays at the center. Until detailed technical plans and labor responses surface, the clearest reading is that AI will arrive in American air traffic control as an assistant, powerful and pervasive, but supervised by the same people who keep aircraft safely separated today.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
