Several times a year, I am asked to join with other speakers to address aviation professionals, normally in the setting of an annual safety standdown or safety day. Inevitably, during these engagements, I am asked to comment on how the rapid pace of developing technology, specifically in aircraft automation and autonomy, will impact the future of cockpit crews and flight departments. It’s an interesting question and deserves a thoughtful answer since it gets to the very core of the flying profession: what’s required of pilots when it comes to contemporary airmanship?
For well over a decade, the arrival of uncrewed air vehicles, remotely piloted vehicles, drones, or whatever you want to call them has piqued the imagination of everyone involved in the aviation profession. Add to that the hype of AI as a threat to virtually every human-occupied position in aviation and you have an environment of unease and insecurity that is hard to escape, especially if you read our industry’s trade publications and listen to the news.
The idea of coming to work one day to find that your copilot has been replaced by a robotic system is not an uncommonly held fear these days. To those outside of aviation and some of us on the inside, it seems as if none of us will be working in cockpits or support roles by the end of the decade.
Unfortunately, all of this talk about the impending removal of humans has had the effect of distracting many of us from real safety focus areas that need our attention right now. These vulnerabilities in both the operator and the technology are as prevalent today as they were before the widespread adoption of advanced flight guidance systems, including flight management systems, precision satellite-based navigation, and dozens of other digital technologies that have proliferated on the flight deck over the past 30 years.
Sometimes I ask skeptics of the role of the human operator how often aircraft systems fail while airborne with passengers and cargo, how frequently aircraft divert because a crewmember becomes incapacitated, or how many times in a day a well-trained and experienced flight crew acts with great skill and precision to safely land an aircraft that has experienced a unique or rare inflight emergency.
The answer to how frequently these situations occur is, simply, all the time.
Should we be engaged in the discussions that describe a more automated future? Absolutely: ignoring the debate is foolish. The resulting systems will be better if flight crews are involved on the front end.
In my view, however, not just as a longtime industry observer but as a crewmember of large, long-range aircraft, most of our focus should be on what will likely be many more decades of “humans in the loop”—and two humans at that—and how they are uniquely suited to use these highly complex systems to deliver increasingly higher levels of efficiency and widening safety margins.
I read several well-respected trade publications every month and browse the most reputable safety websites almost daily. Recently, I was persuaded even more in my opinion of the primacy of the human operator in the cockpit, the cabin, under the wing, and on the shop or hangar floor when I read Missy Cummings’ article in IEEE Spectrum, titled “What Self-Driving Cars Tell Us About AI Risk.”
Cummings is a roboticist who once flew fighter aircraft aboard U.S. Navy aircraft carriers. Her “5 Practical Insights” regarding AI and autonomous systems have applications well outside of surface transportation. Understanding these concepts provides great insight into the importance of our individual “human-in-the-loop” role.
Practical insights regarding AI and autonomous systems:
1. Human errors in operation get replaced by human errors in coding.
2. AI failure modes are hard to predict.
3. Probabilistic estimates do not approximate judgment under uncertainty.
4. Maintaining AI is just as important as creating AI.
5. AI has system-level implications that can’t be ignored.
If you re-read this list and replace “AI” with “flight deck automation,” you can see the close relationship and the influence of these concepts in aviation. For emphasis, consider this quote from Cummings in the journal article:
“Neither the AI in LLMs [large language models] nor the one in autonomous cars can ‘understand’ the situation, the context, or any unobserved factors that a person would consider in a similar situation. The difference is that while a language model may give you nonsense, a self-driving car can kill you.”
For years, as a proponent of wise and thoughtful integration of the human operator with complex technology, I have accumulated a few insights that are eerily similar to those of Cummings. They map over to the above list and come from decades of experience reading, writing, and experiencing firsthand the safety demands of modern aviation. Here are just five of them:
1. Complex systems fail in complex ways (from "Normal Accidents" by George Perrott).
2. The real world happens. Just ask the crews of the most notorious aviation accidents of the past 20 years—those with both “heroic” and “tragic” outcomes.
3. It’s impossible to “automate” all aircraft systems to handle every nuance of contemporary flight operations.
4. The more complex systems there are on the aircraft, the more maintenance and oversight are necessary to ensure smooth and reliable operations on the line.
5. We tend to overestimate technology in the short term and underestimate it in the long run (paraphrased from Kevin Kelly, former editor of Wired magazine).
So where does that leave us, and what practical knowledge can you take from this short discussion? I think it all matters and deserves contemplation—but if it can be boiled down to just one thing, I suggest that, wherever your occupation fits in the broad spectrum of aviation, there are interactions with technology and automation that are worthy of your study and mastery today. Less concern about an imagined, distant future and more focus on becoming the very best professional you can be will yield not only broader safety margins for your flight department but much better personal performance and satisfaction on the job. In my next blog in this series, I’ll address some of the pressing needs surrounding airmanship in contemporary flight operations. Until then, fly safely, and always, fly the airplane first.
Chris Lutat is managing partner of Convergent Performance, a B777 captain, and co-author of “Automation Airmanship: Nine Principles for Operating Glass Cockpit Aircraft.”
The opinions expressed in this column are those of the author and not necessarily endorsed by AIN Media Group.