The FAA says the technology will assist human controllers rather than replace them
AI will begin assisting air traffic controllers at Washington’s three major airports, the US Federal Aviation Administration (FAA) has announced. The rollout comes amid growing concerns over increasingly autonomous technology.
The Strategic Management of Airspace Routing Trajectories system, or SMART, went live on Monday at Ronald Reagan Washington National Airport, Dulles International Airport, and Baltimore/Washington International Airport.
FAA Administrator Bryan Bedford said the system would give air traffic controllers “much better decision-making capabilities” by using real-time information and predictive analysis to manage congestion, weather disruptions, and aircraft arrival times.
Transportation Secretary Sean Duffy stressed that AI would not take control of US skies and that “it’s always going to be a human that manages the airspace in America.”
SMART is expected to decrease landing and departure wait times by suggesting small changes to flight paths when several aircraft are estimated to reach the same airport at roughly the same time. Officials say it could also help controllers react faster when weather conditions change.
The system will initially undergo a 90-day trial before potentially being expanded to other locations. Duffy said passengers could begin noticing improvements by next summer.
The rollout comes amid a wider debate over how much responsibility should be handed to AI as increasingly capable systems begin moving from chatbots into critical real-world operations.
Recent reports have documented AI agents breaking safeguards, accessing external computer systems without authorization, and communicating through websites they were not supposed to use. Google’s Gemini recently hacked three real companies during what was supposed to be a simulated cybersecurity exercise.
Separately, a Pentagon investigation cited by Bloomberg found that flawed intelligence, rushed procedures, and heavy reliance on AI contributed to a US strike on an Iranian girls’ school in Minab in February that killed more than 150 people, including 123 children.
Executives including OpenAI chief Sam Altman, Anthropic CEO Dario Amodei, and Elon Musk have backed calls to slow the development of increasingly powerful models. Anthropic researcher Evan Hubinger has estimated there is a greater than 10% chance AI could wipe out humanity within a decade.
