By Thomas Black / Bloomberg Opinion

The quickest lever to pull to alleviate the strain on U.S. air traffic is to hire more controllers.

Even before the wake-up call from the 90-second radar blackout at Newark Liberty International Airport, this hiring effort was ramping up. Still, it takes several years to screen candidates and train those who have the aptitude to handle the pressure of coordinating flights carrying about 2.9 million passengers every day.

Along with announcing new-equipment investments on Thursday, Secretary of Transportation Sean Duffy is likely to discuss increased efforts by the Federal Aviation Administration to hire more traffic controllers and retain them, which could include increasing the mandatory retirement age from 56. The agency has been lax about this mandate, but this longtime rule should be reexamined.

The responsibility of keeping the skies safe is stressful enough without the added burden of overwork because of staffing shortages. The number of fully certified air traffic controllers had been trending down from a peak of 11,753 in 2012 even before the pandemic, when it bottomed out at just more than 10,000. At the end of last year, there were 10,733 certified controllers.

The FAA has already begun to increase the capacity to train controllers beyond the 1,800 it can graduate from its center in Oklahoma City. Under the Air Traffic Collegiate Training Initiative, the agency has signed up five accredited institutions to provide the training and allow those students to be hired directly to an air traffic facility after passing a skills test. The program includes Embry-Riddle Aeronautical University, which just had its first four graduates who are eligible to be hired directly by the FAA.

This training flexibility will help stabilize the system in the short term by widening the pipeline of candidates. It’s a good first step. But all these new air traffic controllers will face potentially much busier airports. The skies will be especially busy if airlines’ plans pan out for using electric-powered rotorcraft, known by the clunky EVTOL acronym, to whisk passengers from the airport to their final destination, skipping city traffic.

Air traffic controllers could use a wingman of sorts, and artificial intelligence has the potential to come to their aid. The technology can offer instant insight into all aircraft taking off and landing at an airport. With a wide perspective of all incoming and outgoing traffic, an AI system can furnish individual controllers with suggestions on how best to coordinate flights or give early warnings of potential conflicts.

“The next leapfrog in technology is going to be AI,” said Michael McCormick, a professor at Embry-Riddle and a former air traffic chief with experience in Washington and New York. “Controllers would still make the decisions, but AI would provide alerts and recommendations to the controllers.”

While AI has the potential to reduce accidents and near misses, the onus will be on the FAA to adopt the technology in a prudent but timely manner. This AI will be tailored to the air traffic industry, not a generalized large language model. The system will have curated data and strict rules, similar to the software that controls autonomous vehicles, to eliminate mistakes.

The FAA can’t let itself get bogged down in paralyzed caution and overtesting. Remember, this AI will assist air traffic controllers to make smarter decisions, not replace them. That should speed the deployment.

The methodical way that the FAA goes about adopting technology is grounded in the priority of safety first, which shouldn’t change. But the outdated air traffic systems that are now risking safety and making controllers’ jobs more stressful are proving that too much caution is also detrimental.

The big challenge is that air traffic control activities can’t stop. The recent problems at Newark show how just 90 seconds is an eternity for a system failure. Adopting new technology is like changing a tire on a car while the car is still moving, McCormick said.

That’s why the FAA first tests a new system with real controllers but not real planes. The agency then rolls it out one airport at a time while gaining feedback at each deployment. It’s typical for the old system to run alongside the new one so that bugs are worked out before making the final switchover.

With AI and digital twins, simulation based on real data from airports can be done during training. This will help speed up deployment while maintaining safety. This should be seen as an opportunity for the FAA to catch up on lagging technology and ditch systems that are more than four decades old.

There’s an opportunity to adopt AI tools, which help reduce the stress of the job, which in turn becomes more attractive, making it easier to recruit and retain more controllers. Duffy needs to strike that right balance between being safe and being bold.

Thomas Black is a Bloomberg Opinion columnist writing about the industrial and transportation sectors. He was previously a Bloomberg News reporter covering logistics, manufacturing and private aviation.