Taking place at the Flight Deck, moderator Robin Riedel, partner at McKinsey & Company, led experts in the discussion, who agreed AI has a place on business aviation in general, but differed in their expectations for its application in the cockpit.

Further, panelists said this technology is not some futuristic dream – some of these technologies are being applied now.

Mark Groden, CEO of Skyryse, opened the session by explaining what AI is, saying, “AI is as broad of a term as saying ‘software,’” adding printer troubleshooters were an early example of AI.

He explained there are two types of AI – deterministic and non-deterministic. Integrated fly-by-wire is one example of deterministic AI, using “if/then” causal statements with nested equations that govern how the aircraft flies through the sky.

Groden said there are good applications for non-deterministic AI in business aviation, but not necessarily on the aircraft; rather, it has applications related to maintenance, scheduling and other non-flight purposes.

Dr. Yemaya Bordain, president of the Americas, Daedalean AI, which develops technology for optionally piloted and eventually autonomous operations, believes non-deterministic AI will have a place in the aircraft. Bordain said Daedalean is tackling the hard part of how that technology gets certified.

“We do believe there is a place for non-deterministic AI in the aircraft,” said Bordain, citing traffic identification as an example. “A system like ours can identify birds before a strike occurs. Search-and-rescue services are already using our systems to avoid wires. We bring situational intelligence. We’re basically a second set of eyes for the pilot.”

Jon Damush, president and CEO at uAvionix Corporation, said, “The most non-deterministic thing in the aircraft is a pilot at the controls,” adding simplified pilot operations is where AI can improve aviation safety.

For example, uAvionix is using AI to moderate specific areas with unpiloted aircraft in low altitude noncontrolled airspace. The company is deploying a set of classifiers using machine learning to determine what type of items an aircraft might encounter at low altitude – from birds and balloons to ultra-light aircraft.

Robert Rose, CEO of Reliable Robotics, explained the most common causes of small aircraft accidents – controlled flight into terrain and loss-of-control inflight – can be mitigated by AI technology that exists today.

“If continuously engaged autopilot is always on and extremely difficult to turn off, loss of control is no longer a problem. It’s impossible to get the aircraft out of the envelop,” Rose said.

Other examples include new detect-and-avoid technology, which gives pilots not just traffic alerts but also guidance to avoid that traffic. Meanwhile, auto landing systems help prevent runway excursions and taxi systems mitigate the risk of runway incursions.

Getting the technology certified for regular commercial use is one of the biggest challenges, Rose explained, but Reliable Robotics anticipates these technologies culminating in certificated remotely piloted Cessna Caravans in 2027.

Cybersecurity is a priority all experts on the panel said, agreeing that the U.S. should lead in this area. Rose suggested industry standards and community consensus building are the best ways to ensure consistent cybersecurity requirements and encouraged attendees to advocate for U.S. leadership in AI applications in aviation with their legislators.