The $50 Billion Bloodbath: Why Airline Operations Control Centers Will Be Ghost Towns by 2030

[scramjet]

Very fascinating discussion here. Ppragman, interesting discussion on post scarcity. I too am hopeful that the positive of AI will be a fundamental rethink of the economy. In the spirit of full reflexivity, I am biased by lack of appreciation for the cold calculus of profit and loss that seems to run a large part of our social environment. Especially in the face of at least two thousand years of utilitarian-consequentialist thought and progress. People act like money is a concept handed down by God from on high; a part of the fabric of the universe like gravity or something. Same for political ideology, much which has only been shaped in the past 20 years.
I felt like I saw the handwriting on the wall for automation in aviation in the mid-00s with some of the GA products that entered the market. The reality probably has been that fate was cast in the 60s and 70s with advances in microprocessing. However, AI as a concept of change seems to have arisen extremely rapidly in just the past four years and that has taken me by surprise (as if my personal, non-computer science perspective is the end all of general perspective ha).
I always thought space resources would drive post scarcity. Probably they still will, but AI could play a large part.

 

[ppragman]

I felt like I saw the handwriting on the wall for automation in aviation in the mid-00s with some of the GA products that entered the market.

This is more or less my model for how things will go across the whole economy - we've already phased out flight navigators, flight engineers, etc. and we will continue to phase out stuff where we need to modernize. I suspect that we will probably keep 2-crew for pax operations, but I wouldn't be surprised if we see SPIFR big airplanes too? But there's caveats to that too. Regardless, aviation as an industry has the BEST model I've seen full stop for how to work with AI. This may just be deformation professionelle on my part - but yeah, I think that's one thing we're cognizant of and getting right. The aim should not be to fight the robots, but to adapt and utilize them to maximize the benefit to everyone.

You need to hand fly approaches occasionally - so that probably means you need to write some code occasionally, and type out that email here and there, but you don't need to practice holding altitude in cruise for 8 hours. We need to be promulgating that idea to the outside world. "Sure, have Codex write all your boilerplate code" - that's fine, but you should probably try to write a little fun code for yourself here and there and you should also probably stay sharp by jumping in and coding a function here or there so you don't lose proficiency." Would the AI do a better job? Potentially - just like the autopilot can, but the craftsman ethos of aviation, and the general pride and culture of continual self-improvement that nearly EVERYONE in professional flying has is fantastically applicable to a post-AI world. That's the mindset and culture that we can export to other industries.

We stay proficient at hand flying not because we want to be better than the machine, but because we want to be able to perform when the machine isn't there. We want to be good at our jobs - and that requires sweating here and there.

I always thought space resources would drive post scarcity. Probably they still will, but AI could play a large part.

This is my intuition as well. AI and reusable rockets is the ticket to that futuristic world in my mind. Point-to-point sub-orbital space-flight and asteroid mining means that the whole planet is about 72 minutes away. Suddenly a lot of the logistical problems for getting stuff to remote corners of the world really vanishes. Eventually (hopefully) we move all the polluting industries into space, and we start fixing earth. We could turn this place into a garden if we wanted to. But that's a different sort of discussion. That said, across the board (or bored now that I'm on my soapbox again), we could fix most of the problems that grip our modern world if we apply these tools for good. We just have to want to. I think it'll be a lot easier to want to though after we get through this transitional sort of phase we're in now.

 

[A1TAPE]

Pretty sure I read Netjets uses some sort of AI flight planning tool or something, and it tends to be really bad. Saw someone say it routed an EJA executive's flight right through a line of thunderstorms and they sent it off anyone or something along those lines.

Yes it was routed right through a squall line and the ONLY reason anyone knew it did that was there was an issue with the ATC filing. When someone had to manually intervene to file the flight plan they looked at the planned route and realized it was routed through the storm. Had the automation filed the flight successfully then the last line of defense would have been the crew (hopefully) looking at the routing and the weather and realizing the flight was routed through a thunderstorm. Baring that, the crew would have realized when their radar starting showing a sea of red ahead of them.

 

[karkland]

Let's be real. The crew in this situation wouldn't have been the last line of defense. ATC is also looking at weather movement and finding gaps in the weather where they can.

 

[Pancake]

Let's be real. The crew in this situation wouldn't have been the last line of defense. ATC is also looking at weather movement and finding gaps in the weather where they can.

ATC’s primary responsibility is maintaining separation between aircraft, not weather. Staffing issues means a higher likelihood of task saturation which can lead to events where a plane will get cleared/vectored headlong into hail or some other form of dangerous convection.

 

[gopherdx]

Let's be real. The crew in this situation wouldn't have been the last line of defense. ATC is also looking at weather movement and finding gaps in the weather where they can.

ATC does not look nor care about a flights’ route outside of their airspace as long as the filed route doesn’t go against any required routes. Only time a controller will vector around weather is in their own sector.

You’d also be surprised how many ATC reroutes my flights have received right through squall lines only because the initial route gets them out of their airspace but they failed to look at the route downline.

Again, Swiss cheese. This is why dispatchers, pilots, controllers exist. Take one out of the equation, certain situations get 33% unsafer.

 

[karkland]

ATC’s primary responsibility is maintaining separation between aircraft, not weather. Staffing issues means a higher likelihood of task saturation which can lead to events where a plane will get cleared/vectored headlong into hail or some other form of dangerous convection.

While true, they often take weather into consideration. All I am saying is that ATC is still part of the swiss cheese.

 

[///AMG]

I'd say more often than not, ATC is pretty aware of major wx and also the other 10 guys ahead of you probably already asked for the same deviations. Most reroutes/change to filed routing I've had was specifically to avoid known weather. So I wouldn't say they (ATC) get it wrong that often. Sometimes, I'm sure. That being said, most major route changes occur either on initial clearance for departure considerations, or prior to TOD on arrival with a new hot STAR that avoids wx. Like was said, en route normally is ad hoc. I've occasionally seen it proactively changed with pretty well defined wx avoidance routing, but that doesn't seem to be the norm.

 

[who'swho]

One thing that articles like this seem to keep overlooking is the FAA agreeing to a major policy change. And right now I think their policy is that operational control of a flight can only be shared by a human. I'd be amazed to see them change this anytime soon. I see AI certainly being used as a tool, perhaps reducing OCC workload and causing some job attrition that way, but I'd be amazed if the FAA changes a regulation eliminating a well-established safety position.

It will help reduce workload the same way computers did when they were first introduced to the industry. More flights per dxer but less manual work on each one. Let's not forget everyone forecasted the end of our profession when computers replaced manual flight planning. It absolutely reduced the number of Dispatchers and the same will be true with AI. I wouldn't worry about it if I'm already on the seniority list at a major.