Tech Talk: AI is watching your flight at SEA — here’s why that’s a good thing

The next time you fly in or out of Seattle-Tacoma International Airport (SEA), you might not know the airport’s AI and machine learning platforms are watching you in order to help your flight make it to its gate on time, without stopping. Even if your flight is still hours away from landing.

I was standing inside the ramp tower at SEA, listening to air traffic controllers talk with pilots, and looking at a screen dotted with little digital airplanes, each representing an incoming flight. Some of those little planes had a green tag attached to them.

Samer Tirhi, an airline traffic coordinator at SEA, revealed how the color coordination can help speed up the process for travelers catching their flights at the airport.

“If it’s a green data tag to the right of that aircraft, they are clear to their gates,” Tirhi said. “[Red] is an aircraft that will be holding for a gate or does not have a cleared gate into the airport.”

Flight 618 was tagged with red, which in other words means the flight is going to be delayed getting to its gate, even though it’s not scheduled to land for another three hours.

“You could be as far south as Oregon, or even a bit further south than that, and not only will the airport, but air traffic control, the TRACON, and other significant stakeholders of the airport will already know if you’re going to have to hold or not, based on these AI predictions,” Tirhi said.

SEA predicts delays before they happen

Samer and his colleagues are using AI and machine learning to predict delays and fix them before a plane lands at SEA. Using cameras at each gate, they record and measure 36 different actions ground crews take, chocking the wheels, offloading, on-loading passengers, luggage, cargo, animals, refueling, restocking, and much more.

They can also track birds and vehicles and detect any debris on the runway, then mix all of that new data with historical data and which airline they’re talking about.

“From there, we can aggregate when the aircraft will be off of the gate for the incoming aircraft,” Tirhi said.

Ed Appleberry, who’s in charge of SEA’s ramp controllers, stated that the airport is focused on remaining one step ahead of its constituents.

“Most airports are jealous of what we have up here. They want to do it,” Appleberry said. “We’re trying to stay ahead of the traffic and congestion at Seattle.”

SEA ranks as the 12th busiest airport in the country, and dead last among all medium and large hub airports in surface area. There’s just not a lot of room for the number of planes.

Back to flight 618, it’s on final approach to land, but another plane is still blocking its gate. Machine learning tools instantly predict the delay, giving ramp controllers and the airline time to act.

The airline pulls back the sky bridge and attaches a tug while waiting. Moments later, AI updates, clearing them to send flight 618 in. By the time it arrives, the delayed plane is gone.

“Our motto is safe, orderly, and expeditious flow of traffic,” Appleberry said.

With AI and machine learning, SEA is hitting those goals faster than ever before.

“It feels like every time we talk to an aircraft, we can do a little extra work, a little extra coordination, and find a way to be more efficient and save them two or three minutes of taxi time. That’s what we like to do,” Appleberry said.

Follow Luke Duecy on X. Read more of his stories here. Submit news tips here.