Airline Pilot Central Forums

Airline Pilot Central Forums (https://www.airlinepilotforums.com/)
-   Aviation Technology (https://www.airlinepilotforums.com/aviation-technology/)
-   -   The Future Of Artificial Intelligence (https://www.airlinepilotforums.com/aviation-technology/104285-future-artificial-intelligence.html)

C130driver 07-28-2017 05:15 PM


Originally Posted by madmax757 (Post 2400141)
And you even just said it yourself in your reply. It pushed buttons and got the machine on the ground.

Also I said CO PILOTS in 5-10 years. Not the AC commander or Captain. I fly an A320 and I recall a recent flight. AP1 on at 100 feet, pushed some buttons , talked on radio, visibility below Cat 1 mins so did an autoland. Taxi to gate. Its just too easy.

Didn't mean to dis anything about the C17 mission. I do realize you fly into hostile areas and do tactical approaches where you get shot at and refuel inflight. But don't you also drone from point A to B quite often also? Couldn't you do most missions with one human pilot and have a robot copilot in 5-10 years?

If you re read the article, it stated the Robot could manipulate the auto flight system and get the airplane on the ground if the Captain becomes incapacitated. That's what they are demonstrating.

Don't you as an A320 pilot ever find yourself intervening during managed modes, heading select mode around weather, etc? We have two pilots for 2 reasons: contingency and CRM. A computer is no where near the point to be able to back up one pilot programming the "box." What happens when the pilot is incapacitated? The computer just knows to do what it is programmed.

Now you could argue that the second pilot would be in the ground station, but then how much are we really saving cost wise; considering the investment into the automation and still having to pay that guys' salary. Not much, not to mention the insurance costs and non monetary costs/risks. A cargo plane crashing isn't really cheap. The only true savings would be a pilot less cockpit but like I said that is generations away. AI would have to be developed to a point where it can actually make human like decisions

I don't fly C-17s, just an example, but the C-130J has a similar cockpit. Even on a routine mission there is no way one pilot could take it from start to finish.

atpcliff 07-28-2017 05:21 PM


Originally Posted by NeverHome (Post 2400606)
Not quite where I was going. Yes electric motors, sure, but where are they going to draw their power from? I know batteries! But where will the batteries draw their power?

Consider that 1 gallon of gasoline has approximately 115000 BTU of energy. We have absolutely Nothing that can even come close in energy! Solar energy via panels doesn't even come close, plus there is a time factor involved (think time derivative). So how are we going to power those fancy nasa electronic motors?

This I feel is the real problem! Add to that the overall efficiency of a system. What we desperately need is an alternative fuel source that can at least come close to fossil fuels in both BTU output, and time derivative. The other part is improving the efficiency of engines that utilize this fuel.

Yes if we could find an engine that operates on technology other than thermal energy, that would be good. But short of a nuclear reactor, we do not have this technology (or at least a mobile version).

All is not lost, I think there is real potential in bioengineering of fuels. My vote also goes to oilshale. Neither oil nor shale, but has enormous potential!

My point is this. Before AI comes of age, fossil fuel dependence will be a show stopper. So in the meantime that is a wonderful area to invst in.

My .02 :cool:

6-9 passengers, 600 mile range on one charge:
https://futurism.com/this-all-electr...single-charge/

Mesabah 07-28-2017 09:56 PM


Originally Posted by C130driver (Post 2401041)
Don't you as an A320 pilot ever find yourself intervening during managed modes, heading select mode around weather, etc? We have two pilots for 2 reasons: contingency and CRM. A computer is no where near the point to be able to back up one pilot programming the "box." What happens when the pilot is incapacitated? The computer just knows to do what it is programmed.

Now you could argue that the second pilot would be in the ground station, but then how much are we really saving cost wise; considering the investment into the automation and still having to pay that guys' salary. Not much, not to mention the insurance costs and non monetary costs/risks. A cargo plane crashing isn't really cheap. The only true savings would be a pilot less cockpit but like I said that is generations away. AI would have to be developed to a point where it can actually make human like decisions

I don't fly C-17s, just an example, but the C-130J has a similar cockpit. Even on a routine mission there is no way one pilot could take it from start to finish.

Machine learning systems look at millions of human flight hours, and then copy what humans would do in that situation. It's a completely different system then what is currently installed in aircraft. Humans would be free to make judgement calls, while all flying is done by the autopilot.

Anderson 07-28-2017 10:11 PM


Originally Posted by Mesabah (Post 2401133)
Machine learning systems look at millions of human flight hours, and then copy what humans would do in that situation. It's a completely different system then what is currently installed in aircraft. Humans would be free to make judgement calls, while all flying is done by the autopilot.



So the systems just copy what humans would do? That doesn't sound like AI.



Sent from my iPhone using Tapatalk

samsmug 07-29-2017 09:44 PM


Originally Posted by trip (Post 2397990)
You should get out of the 141 school and into programming/coding ASAP.

Agreed. Airline pilots will be obsolete in 10 years, possibly less. Cargo first, then passenger. It's going to happen faster than we want to admit.

Go learn to write 3d computing, machine learning or deep learning code.

We're talking about career paths on opposing trajectories...good luck whichever way you go.

badflaps 07-30-2017 03:52 AM

Going the way of the single engine airliners, first smoking hole, revision.

CBreezy 07-30-2017 04:01 AM


Originally Posted by samsmug (Post 2401543)
Agreed. Airline pilots will be obsolete in 10 years, possibly less. Cargo first, then passenger. It's going to happen faster than we want to admit.

Go learn to write 3d computing, machine learning or deep learning code.

We're talking about career paths on opposing trajectories...good luck whichever way you go.

Obsolete in 10 years? That's a dumb statement. Airlines and cargo companies are flying equipment that is nearly 30 years old. Even if it took 10 years to develop a fully auto airliner (it will take far far longer), it'll take 20-30 to cycle all the old aircraft out of the system. Current tech isn't even close to being ready. We are probably 40 years away from that possibility.

tomgoodman 07-30-2017 07:23 AM


Originally Posted by badflaps (Post 2401573)
Going the way of the single engine airliners, first smoking hole, revision.

Then they can announce: "We will replace those dangerous robots with human pilots, and save a bundle of money too!" :D

NeverHome 07-30-2017 08:50 AM


Originally Posted by atpcliff (Post 2401043)
6-9 passengers, 600 mile range on one charge:
https://futurism.com/this-all-electr...single-charge/

Very impressive.

2 thoughts though. The first being, where did the electrical energy come from to charge the battery?

The second thought is how economical is the solution you have presented?

Consider this: 1 gal of gasoline is approximately 115000 BTU. So say in an airliner burning 2000 lbs per hour we are burning 33810000 BTU per hour. Thats Alot of energy. In all fairness there is alot of inefficiencies in our heat engine. Even at 30% efficiency or stated another way, if we need only 30% of a gallon at 100% efficiency, than we will need at least 10143000 BTU for this example.

Where are we going to get that energy? Batteries are great but they dont run on PFM. That energy comes from somewhere.

samsmug 07-30-2017 09:37 AM


Originally Posted by CBreezy (Post 2401575)
Obsolete in 10 years? That's a dumb statement. Airlines and cargo companies are flying equipment that is nearly 30 years old. Even if it took 10 years to develop a fully auto airliner (it will take far far longer), it'll take 20-30 to cycle all the old aircraft out of the system. Current tech isn't even close to being ready. We are probably 40 years away from that possibility.

I'll concede that 10 years is likely to be the soonest we'll see automation of cockpits, but I stand by the sentiment of my post. Once a pilot is no longer needed, we are obsolete. Culture and politics will determine how fast we are out of work.

Automation is coming, it may be 10, 20, or 100 years away. Before it's a completed process, our profession may undergo significant changes in terms of QoL and compensation as a result of management's automation leverage.

Given the current trajectory of machine learning, I think the OP is taking on significant risk by investing in a long term airline career. I'd at least recommend having a backup plan in case this happens sooner than later.


All times are GMT -8. The time now is 05:06 AM.


Website Copyright © 2026 MH Sub I, LLC dba Internet Brands