Retired airline pilot saved over 400 lives after an engine explosion midair. He says AI could make the job harder, not easier.

• In 2010, Qantas Flight 32 faced engine failure shortly after takeoff, causing a crisis aboard.
• The pilot in command and his crew saved everyone by knowing when to follow and ignore the computers.
• He told Business Insider that AI and automation make flying harder, especially in times of crisis.

When Qantas Flight 32 left Singapore on November 4, 2010, everything felt routine. It was clear skies, no wind, the "perfect day to go flying," recalled Captain Richard Champion de Crespigny, the pilot in command.

That didn't last long. Four minutes after takeoff, two loud bangs tore through the cabin, shaking the aircraft as dozens of warning alarms erupted.

"The engine number two had exploded, and it created shrapnel like a cluster bomb where 400 pieces of shrapnel hit the aircraft," de Crespigny told Business Insider's Maggie Cai, who interviewed him for our video series, Authorized Account.

De Crespigny, a former Royal Australian Air Force pilot, led a five-person crew that day. They faced 21 system failures, 120 checklists, 650 broken wires, 50% network failures, and shrapnel holes through the wing.

Yet, every one of the 469 people on board survived, thanks in part to de Crespigny and the crew knowing when, and when not, to trust the computer systems that were failing around them.

A later investigation found that the flight failed due to a manufacturing defect in a small stub pipe in the Airbus 380's Rolls-Royce engine. The threat that worries de Crespigny these days, however, is automation.

"Whether we like it or not, automation is here and it's going to control our lives more every day in aircraft," he said. "Automation presents more problems for pilots, not less," he said.

That's because when black-box computers fail, and they do fail, he said, then pilots must have the ability to identify the failed computers, disconnect them, and then manually fly the aircraft.

Computer override

After the explosion on Flight 32, the aircraft's digital monitoring system flooded the screen with checklists, de Crespigny said, but he and the crew didn't just blindly follow them.

"Some of the checklists we did, no question. Some of the checklists we queried, and we still did it, but there are quite a few checklists that if we had completed them, it would've changed the outcome of the flight. So we ignored them," he said.

For about two tense hours, de Crespigny and his crew circled above Singapore, methodically checking each system to understand what had failed and what still worked before bringing the plane safely back to land at Singapore airport.

De Crespigny has since written two bestselling books about his experiences as a pilot, launched and hosts his FLY! podcast, and gives presentations worldwide on resilience across both personal and corporate topics.

He retired from flying in 2020 but remains uneasy about AI and automation, saying that "we are more of a servant to the box."

"We have to disrupt ourselves and keep control because if we don't, we will be disrupted and we'll have no control," he added.

This is especially true for crisis situations, like what happened on Qantas Flight 32. With increasing dependency on AI and automation, he said that pilots must commit to a lifetime of learning, using AI and automation only as tools, and retain their manual piloting skills, which are necessary in crisis situations.

The future of flying

While AI disrupts jobs across industries, AI replacing pilots is one thing that de Crespigny is not worried about.

"AI, as it stands today, is not threatening pilots in the cockpit. AI is a tool, it's not a replacement," he said, adding that, "Until we see sentient computers in the workplace, there will be pilots in aircraft."

However, he does see autonomous planes as a possibility in the distant future.

The first autonomous pilotless aircraft will likely be used by the military and for overwater cargo operations, he said.

In about 30 years, though, we could have the technology to fly a passenger plane autonomously with sentient AI, he predicted.