FAA seeks predictive AI
- The FAA is accepting bids for SMART, an AI system to help forecast air-traffic routing conflicts up to two hours ahead. - SMART is part of a broader $32.5 billion modernization plan that includes replacing hundreds of radars and expanding staff. - Companies such as Palantir, Thales and startups are reportedly competing to build the forecasting and decision-support system (upi.com).
The Federal Aviation Administration is taking bids for software that would warn controllers about flight-path conflicts up to two hours before they happen. (upi.com) Transportation Secretary Sean Duffy said on April 18 that today’s tools often flag trouble about 15 minutes ahead, while the new system would let controllers make small route changes 90 minutes to two hours earlier. He said the goal is to reduce bottlenecks and scheduling conflicts at busy airports. (upi.com) The project is called Strategic Management of Airspace Routing Trajectories, or SMART. UPI reported that Palantir, Thales and Air Space Intelligence are bidding, and that the Department of Transportation and the FAA expect to give a progress update on April 21, 2026. (upi.com) Air traffic control works by separating aircraft in time and space, and controllers now rely on radar, weather feeds and automation to spot conflicts as flights move from airport areas to high-altitude routes. The FAA said in November 2025 that it wants newer systems that can move flight data between facilities more efficiently and help controllers organize airspace with fewer delays. (faa.gov) SMART sits inside a much larger rebuild of the national system. UPI reported the package at $32.5 billion, and the FAA said in June 2025 that the overhaul would replace core radar, software, hardware and telecommunications networks. (upi.com) (faa.gov) The FAA has framed the broader push as a response to an aging system and rising complexity in the National Airspace System, the federal term for the network of airports, routes, towers and control centers. In June 2025, Duffy said the existing setup was “showing its age,” and the agency began seeking an outside integrator to help build a replacement. (faa.gov) The agency has also been laying out rules for how artificial intelligence should be used in aviation. In its AI safety roadmap, the FAA said AI can support safety but must be shown to be safe before it is put into use, because systems that learn behave differently from traditional software built only by fixed design. (faa.gov) That makes SMART less like an autopilot project than a forecasting tool for controllers on the ground: software that looks farther ahead so humans can act sooner. The next test is whether the FAA can turn that idea into a working system inside a modernization effort it has already started to build. (upi.com) (faa.gov)
