AI and aircraft (spun off from Fedex thread)

There is no world in which any pilot is paid too much. The sacrifice to get into this profession, the sacrifice to maintain this profession, and the risks involved in this profession require a compensation level that allows us to be able to come to work as fully rested and mentally ready to work as humanly possible. People oversimplify this job, but if you happen to have a good day… cool… That’s not what we’re paid for… We’re paid for all the garbage that can and will go wrong on an almost daily basis. We stop the swiss cheese holes from lining up so often people take it for granted and forget what a pilot is really worth…

Junior pilots with inexperience… I’m not sure in what world taking compensation from senior pilots to give to junior pilots to make them “better pilots” is a reasonable solution. Whatever it takes to make a pilot safe is what it takes, and has nothing to do with taking money away from people who have earned and deserve it. Right now in aviation we’re seeing a lot of issues related to rapid hiring and growth that compromise safety. There have been various threads about this issue on this forum. A true solution really hasn’t been found as what it takes to truly be a “good/safe” pilot can not be fully defined by simply meeting training requirements. The best we can do is mentor pilots along the way, and find those with good skill and character to move forward in this profession, and unfortunately at the end of the day there just isn’t much substitute for the almighty experience needed to be effective, proficient, and safe. Also it is difficult to mentor and/or find the ones with good skill and character when people that lack one or both those traits are easily allowed through this maze of a career as hiring standards, interview requirements, and flight experience requirements decline… I don’t have a good solution to solve these issues. Everyone needs to watch out and take care of each other more now than ever to get the job done safety, but throwing more money at this issue isn’t going to fix things…



Sorry, not buying it.


We're overpaid now - to the point that it's worth while for airlines and the industry in general to figure out pilotless airliners. AI will be taking over our jobs at some point in the future.


And ALL that talk about being the last line of defense, that garbage that can go wrong on a daily basis, will quickly end when Boeings and Airbus start flying pilotless around the world without any issues - and actually improve safety because they make pilot-error crashes non-existent.


While I'm happy with the salary I make, I realize we're contributing to making pilotless airliners happen sooner. If I was an airline CEO, I'd want pilots gone. 400K+ salaries, age 65 retirement issues, sick calls, fatigue calls, pickets, strikes, etc.






You know who won't have those issues?


AI.
 
Sorry, not buying it.


We're overpaid now - to the point that it's worth while for airlines and the industry in general to figure out pilotless airliners. AI will be taking over our jobs at some point in the future.


And ALL that talk about being the last line of defense, that garbage that can go wrong on a daily basis, will quickly end when Boeings and Airbus start flying pilotless around the world without any issues - and actually improve safety because they make pilot-error crashes non-existent.


While I'm happy with the salary I make, I realize we're contributing to making pilotless airliners happen sooner. If I was an airline CEO, I'd want pilots gone. 400K+ salaries, age 65 retirement issues, sick calls, fatigue calls, pickets, strikes, etc.






You know who won't have those issues?


AI.
1683004394745.gif
 
You know who won't have those issues?


AI.
Nope.

And dude, how are we ‘overpaid’ - living like we do in SoCal. Where a 1100 sq ft. starter house is $600k and was built before anyone heard of the Beatles.

Oh, look, that’s weird - there’s an elephant in the sky. Then the AI flies right into a TS.

Or through moderate turbulence.

Or it hits mountain wave and ends up in the coffin corner because it’s flying at Max Alt. to save 30 lbs of gas.

The tech still isn’t there to replace us. And the public probably won’t sit still knowing that it’s a robot up front.

We can’t even eliminate gridlock traffic at airports and somehow the planes are gonna fly themselves.
 
Nope.

And dude, how are we ‘overpaid’ - living like we do in SoCal. Where a 1100 sq ft. starter house is $600k and was built before anyone heard of the Beatles.

Oh, look, that’s weird - there’s an elephant in the sky. Then the AI flies right into a TS.

Or through moderate turbulence.

Or it hits mountain wave and ends up in the coffin corner because it’s flying at Max Alt. to save 30 lbs of gas.

The tech still isn’t there to replace us. And the public probably won’t sit still knowing that it’s a robot up front.

We can’t even eliminate gridlock traffic at airports and somehow the planes are gonna fly themselves.


I'm not saying you and I are getting replaced. We'll end our careers fine - probably.


But our kids? I'm not sure I'd push/encourage mine into flying.
 
Sorry, not buying it.

We're overpaid now - to the point that it's worth while for airlines and the industry in general to figure out pilotless airliners. AI will be taking over our jobs at some point in the future.

While I'm happy with the salary I make, I realize we're contributing to making pilotless airliners happen sooner. If I was an airline CEO, I'd want pilots gone. 400K+ salaries, age 65 retirement issues, sick calls, fatigue calls, pickets, strikes, etc.

You mean the same issues airlines have been dealing with for the last 60+ years?

Look at airline pilot pay back in the 1980s and adjust them for inflation… Airline pilots have been getting paid the same more or less all along, and in a lot of cases actually a lot less for quite a while.


And ALL that talk about being the last line of defense, that garbage that can go wrong on a daily basis, will quickly end when Boeings and Airbus start flying pilotless around the world without any issues - and actually improve safety because they make pilot-error crashes non-existent.

You know who won't have those issues?

AI.

Putting computers on a plane just moves the last line of defense from the air to the ground increasing risk significantly. The safety compromised by doing that now is pretty risky and expensive. Maybe someday, but for today, tomorrow, and the foreseeable future pilots are here to stay on transport category aircraft. Don’t kid yourself. I’m sure you’ve found yourself in plenty of situations that AI wouldn’t have been able to figure out… I know I have…
 
... I’m sure you’ve found yourself in plenty of situations that AI wouldn’t have been able to figure out… I know I have…
While I want to agree with this thought, I suspect that the near-future of AI will surprise us with its capabilities. I can foresee AI having a greater pool of possible responses/actions available than a human will have ever experienced, and the ability to very quickly analyse the outcome of all those possible responses to a situation...
 
I am always perplexed by people who think that AI is incapable of matching human thought. What exactly is it about a computer made of meat that you think makes it inherently superior to a computer made of microprocessors?
 
While I want to agree with this thought, I suspect that the near-future of AI will surprise us with its capabilities. I can foresee AI having a greater pool of possible responses/actions available than a human will have ever experienced, and the ability to very quickly analyse the outcome of all those possible responses to a situation...

For pure logic based “AI”, like programming I do think it’s possible near term.

I’ve messed around with ChatGPT and GoogleAI quite a bit. It’s very “confidently incorrect” with a lot of subjects. That’s not exactly a robust system for safety related operations.
 
For pure logic based “AI”, like programming I do think it’s possible near term.

I’ve messed around with ChatGPT and GoogleAI quite a bit. It’s very “confidently incorrect” with a lot of subjects. That’s not exactly a robust system for safety related operations.

This will be the true Achilles heel of AI. It’s like a child, it only knows what it has been told, and the system makes no allowances for false information, bad actors or plain old BS.

At some point, there will be some nasty example of AI being completely wrong because of incorrect information being fed to it. On the surface, it will have seemed to be wrong, but “the computer said so”, and the normal cross check will be ignored, and this will be the driving force behind some massive wreck, building collapse or some other similar disaster.

To be fair, I’ve seen quite a few non-artificially intelligent people in the sciences, engineering and other related fields be “confidently incorrect”, sometimes massively so, and some actually chose to do so for financial and/or reputational gain, so nothing new there.
 
For pure logic based “AI”, like programming I do think it’s possible near term.

I’ve messed around with ChatGPT and GoogleAI quite a bit. It’s very “confidently incorrect” with a lot of subjects. That’s not exactly a robust system for safety related operations.

So you’re saying…


It mimics human behavior well.
 
Nope.

And dude, how are we ‘overpaid’ - living like we do in SoCal. Where a 1100 sq ft. starter house is $600k and was built before anyone heard of the Beatles.

Oh, look, that’s weird - there’s an elephant in the sky. Then the AI flies right into a TS.

Or through moderate turbulence.

Or it hits mountain wave and ends up in the coffin corner because it’s flying at Max Alt. to save 30 lbs of gas.

The tech still isn’t there to replace us. And the public probably won’t sit still knowing that it’s a robot up front.

We can’t even eliminate gridlock traffic at airports and somehow the planes are gonna fly themselves.

I think AI will know not to fly at coffin corner or fly into anything looking funky.

They'll teach it by feeding (I dunno) millions of hours worth of flying into it. AI will see that we go around anything that looks like an odd cloud. It'll see that an odd cloud at night illuminated slightly by the moon plus some green on the radar, the pilots deviated every time....

Max altitude? Pilots never went there when it got bumpy or the charts showed potential for mountain wave.

Just expose it to enough and it'll learn.
 
I am always perplexed by people who think that AI is incapable of matching human thought. What exactly is it about a computer made of meat that you think makes it inherently superior to a computer made of microprocessors?

I think it’s the 5 senses and our abillity to use them and AI isn’t there yet. Can AI detect when other pieces of the puzzle are not put together correctly? Failed type IV fluid on a snowy day? An electrical fire. A rejected takeoff and evacuation? A flight control binding or rigged backwards? A wheel bearing failure? Grabbing brakes? Underserviced parts making terrible grinding noises (needing grease and/or inspection). CPDLC, HF, or VHF glitches or hick ups? Burnt out lights? Take the entire sum of maintenance write ups in an entire fleet, and find what percentage of those AI could solve of course throwing out the things AI no longer needs like O2 or cockpit panel lights.

I think we have a long way to go… It’s not necessarily human thought, but human senses. Although I do believe AI is still far off from human thought required in aviation. You need a relatively sterile environment, and aviation is just not that sterile even as sterile as it’s been made to be…
 
I agree that AI isn't ready...yet. While the practicality of retrofitting the upcoming technology to existing equipment will probably be a rather long-term process, the ability of AI to do the things that we do is, I'm guessing, not that far off.

Ability to replicate human sensory input? Force sensors on flight controls for feedback. Vibration sensors throughout an aircraft that can isolate engine vibrations from wheel vibrations from aerodynamic vibrations caused by loose or damaged skin, etc. Visual systems that can look at clouds, and let AI compare it to millions of images that it has catalogued along with their corresponding threat profiles, and cross check visuals with digital radar return inputs. Visual systems looking at all critical flight surfaces and comparing them to millions of images in order to monitor status on an instantaneous basis. Feed AI hundreds of thousands of photos of what Type IV looks like under different lighting conditions and at all stages of pre-failure and failure, and it will become better at determining when to call for inspection or reapplication than your average line pilot.

I could go on and on, but the bottom line is that if we can give AI proper information inputs (better and more detailed than a human can take in), give it all available courses of action, teach it how those actions affect conditions, teach it the full scale of desirable outcomes versus undesirable, and eventually AI will make better decisions, quicker than a human can.
 
I agree that AI isn't ready...yet. While the practicality of retrofitting the upcoming technology to existing equipment will probably be a rather long-term process, the ability of AI to do the things that we do is, I'm guessing, not that far off.

Ability to replicate human sensory input? Force sensors on flight controls for feedback. Vibration sensors throughout an aircraft that can isolate engine vibrations from wheel vibrations from aerodynamic vibrations caused by loose or damaged skin, etc. Visual systems that can look at clouds, and let AI compare it to millions of images that it has catalogued along with their corresponding threat profiles, and cross check visuals with digital radar return inputs. Visual systems looking at all critical flight surfaces and comparing them to millions of images in order to monitor status on an instantaneous basis. Feed AI hundreds of thousands of photos of what Type IV looks like under different lighting conditions and at all stages of pre-failure and failure, and it will become better at determining when to call for inspection or reapplication than your average line pilot.

I could go on and on, but the bottom line is that if we can give AI proper information inputs (better and more detailed than a human can take in), give it all available courses of action, teach it how those actions affect conditions, teach it the full scale of desirable outcomes versus undesirable, and eventually AI will make better decisions, quicker than a human can.

I’ve posted about this before. One of the top execs at Microsoft gave me a white paper that showed AI was 15 years away from being able to replace heavy equipment operators. Sorry that’s what we do so there ya go.

I don’t dispute that.

The discussion that took place when I pointed out that our society can’t manage the kind of infrastructure update that would make it possible for AI to control an airplane with passengers inside, was pretty interesting.

I still keep saying this and exactly what you’re saying over and over. The only responses I get are usually condescending remarks from Todd and CC.

Which brings us back to the Saturday Zoom discussion among form members on the falling value of JC as a community.


Sent from my iPhone using Tapatalk
 
For pure logic based “AI”, like programming I do think it’s possible near term.

I’ve messed around with ChatGPT and GoogleAI quite a bit. It’s very “confidently incorrect” with a lot of subjects. That’s not exactly a robust system for safety related operations.

This is where I think a lot of confusion comes from. ChatGPT is not AI. It’s relatively basic machine learning. True AI is adaptable and ”thinks“ for itself. It’s coming.

I think it’s the 5 senses and our abillity to use them and AI isn’t there yet.

Oh, I agree it’s not there yet. But it’s fast approaching, and I see a lot of wishful thinking from people that it isn’t even possible.
 
I’ve posted about this before. One of the top execs at Microsoft gave me a white paper that showed AI was 15 years away from being able to replace heavy equipment operators.

Maybe. But if you go back to integrated circuit development prior to 1965 the idea of density doubling every 2 years wasn't even considered. 60 years later, Moore's Law (stated in 1965) has held up. We may be at the prior 1965 point of AI, or we may not. But I certainly wouldn't put my full faith in a whitepaper from more than two years ago, in a field that moves as fast as this one is.
 
This is where I think a lot of confusion comes from. ChatGPT is not AI. It’s relatively basic machine learning. True AI is adaptable and ”thinks“ for itself. It’s coming.



Oh, I agree it’s not there yet. But it’s fast approaching, and I see a lot of wishful thinking from people that it isn’t even possible.

Perhaps. But I think AI controlled airliners are absolutely NOT the first target in the crosshairs of AI.
 
Back
Top