The US Federal Aviation Administration is deploying SMART, an artificial intelligence tool designed to help manage air traffic congestion in the Washington, DC airspace, with a potential nationwide rollout to follow. Expected to launch as soon as 21 September, the system will advise air traffic controllers by predicting traffic flows and identifying potential conflicts based on factors such as airline schedules, weather, and airport capacity. This phased approach is considered prudent by aviation experts, given the complexity and unknowns inherent in deploying such a large-scale AI system across America's entire national airspace.

The SMART system is part of an $875 million, 12-year contract awarded to Boston-based Air Space Intelligence, which also covers development of a new Flow Management Data and Services system to replace the current air traffic control system in Virginia. Rather than changing existing procedures, SMART will provide "alternative route information" through existing FAA systems, a distinction that reassured the airline industry after initial concerns about the deployment. The system aims to reduce aircraft fuel burn, improve flight punctuality, and expedite recovery from weather disruptions and congestion, whilst operating initially on aircraft flying above 24,000 feet, though questions remain about which types of AI models—deterministic, machine-learning, or others—are being employed.