Boeing looking into passenger planes without pilots

All I can say is good luck qualifying your AI or neural net to DO-178B class A having had experience with class B software.
 
That is how I got it from the air crash investigation episode regarding the scenario.

The final part of the flight after the right turn over DXB was mostly uncontrolled. By this time it was also a one man operation as the captain was more than likely already incapacitated, or dead, somewhere outside of the cockpit.
 
Let's be honest here: these companies are saying that cost of implementing smarter systems is more than the cost of potential lost lives. I have a massive problem with that, but apparently regulatory authorities don't.

It is cost vs benefit. Regulatory authorities understand it, but you apparently don't.

It isn't going to help anyone if regulatory authorities push stuff onto airlines that makes operating them so expensive that only the super rich can afford them.
 
It is cost vs benefit. Regulatory authorities understand it, but you apparently don't.

It isn't going to help anyone if regulatory authorities push stuff onto airlines that makes operating them so expensive that only the super rich can afford them.

Like a DO-178B class A or equivalent software
 
Garbage in, garbage out. If you have incorrect information, you're probably going to make incorrect decisions. Had the instruments (sensors) been more useful and indicated the actual error, a human or AI pilot would have been able to make a better decision that could have saved the plane.

Then let me sing the same sing song again... The sensors aren't always sophisticated enough to feed the correct data and the systems not smart enough nor reliable enough to interpret it.

As has been said in this thread before, flight computers often go "ok, I don't know what I'm doing anymore - help please." and then the pilot has to intervene. When there's no human onboard, then what?

Look, this is fairly basic. Nobody is saying AI will never take over from human pilots. The thing is, the tech just isn't anywhere near ready yet. A full axis autopilot is one thing, full-scale autonomous AI is another.
Again, look at Qantas Flight 32. There were so many highly critical errors which demanded attention, AI would never have found the actual issue. In fact, AI hampered the pilots in determining what the actual issue was.

Again, on type, you see that on 40% of flights you see the computers screw up - in one way or another. I don't get why this is ignored. Now, you won't see Airbus crapping themselves over this because there are pilots onboard who compensate for this - most of the time the compensation is inadvertent. The popular line among Airbus pilots are "and what is she doing now?" It's an accepted part of flying. Just the other day, a friend of mine had an issue with a hydraulic leak which caused the flight computers to lose their mind. In the end, you have to fight the computers every step of the way in terms of getting the aircraft back on the ground. Then we saw the Air Asia flight which decided it wasn't going to fly to the destination, instead, the flight computers coupled the autopilot and had it fly to a completely different destination altogether - there wasn't anything the pilots could do about it... and now with the flights in and out of St. Maarten. The airport is badly damaged and not exactly serviceable for airline ops, but I just got a message from a friend who will be taking his operator's A320 there to ferry some survivors out. The computers have an expected approach and expected landing type, but it goes without saying that these procedures will not and cannot be followed in this case.

And then there's the issue of Africa and the third world. Take Ndola airport (don't feel like wiring out it's full name) for example. The AIP states that runway lights don't officially exist in Zambia, but the Jeppesen charts indicates there is. No NOTAMs exist which details the recent work which was done in removing some of these lights (randomly it seems) or that landing on the centreline would result in damage to aircraft. FOD exist in certain parts of the airport, again, none of it documented. The runway has some patches and some holes, with no work in progress ever being documented. So you've got first world safety standards and first world tech in third world countries - which, by the way, can't be dismissed...the number of third world countries isn't insignificant by any means.

I don't believe that most crashes can be 100% attributed to pilot error (there are usually many factors at play - such as AI screwing up and the pilots not realizing it :p), that's a ignorant statement - one which not even the authorities, in all their ignorance, make. Having tech on board has greatly improved safety over the years (ok, in some degree...the Cirrus SR22 being a prominent exception of this), and will continue to do so. However, for at least the next few decades, the world will not see fully autonomous airliners taking over...single pilot operations? Maybe, but the regulatory bodies of this world will have to catch up first. And let's be sure that they will take their jolly time to do so - single pilot ops are currently way too restrictive. We'll need to see AI capable of handling operations on it's own for at least 98% of the time before those limits will start to change - and we're far away from that mark.
 
Last edited:
I'm gonna sing my AI song here too... :)

I would feel more comfortable for a centrally controlled really powerful AI to fly planes. Not only does it not have psychological problems, human error etc.

But, it can also get input from all the aircraft sensors in milliseconds and decide on the most logical course of action when a problem arises. For human pilots its a case of reading sensors, trying to remember what u studied, processing multiple inputs, manuals etc...

I cannot see how a human is faster at recovery from an error than an artificial intelligence that can get input in milliseconds and respond to all all the necessary components and controls without needing hands and fingers.
 
I'm gonna sing my AI song here too... :)

I would feel more comfortable for a centrally controlled really powerful AI to fly planes. Not only does it not have psychological problems, human error etc.

But, it can also get input from all the aircraft sensors in milliseconds and decide on the most logical course of action when a problem arises. For human pilots its a case of reading sensors, trying to remember what u studied, processing multiple inputs, manuals etc...

I cannot see how a human is faster at recovery from an error than an artificial intelligence that can get input in milliseconds and respond to all all the necessary components and controls without needing hands and fingers.

Not meaning to be harsh here, but did you literally just ignore my entire post? Nobody is arguing about whether AI will EVER be flying autonomous airliners. Now? No, not by a long shot. Soon? I doubt it.

Tech issues cause most drone accidents btw - just as a side note.
 
I'm gonna sing my AI song here too... :)

I would feel more comfortable for a centrally controlled really powerful AI to fly planes. Not only does it not have psychological problems, human error etc.

But, it can also get input from all the aircraft sensors in milliseconds and decide on the most logical course of action when a problem arises. For human pilots its a case of reading sensors, trying to remember what u studied, processing multiple inputs, manuals etc...

I cannot see how a human is faster at recovery from an error than an artificial intelligence that can get input in milliseconds and respond to all all the necessary components and controls without needing hands and fingers.

The day that AI figures out that the best course of action is to land itself on a river or to use asymmetric thrust for directional control is the day I will step onto a pilotless aircraft.

Humans will never be faster than AI but AI can't make s**t up to save itself.
 
Last edited:
I'm gonna sing my AI song here too... :)

I would feel more comfortable for a centrally controlled really powerful AI to fly planes. Not only does it not have psychological problems, human error etc.

But, it can also get input from all the aircraft sensors in milliseconds and decide on the most logical course of action when a problem arises. For human pilots its a case of reading sensors, trying to remember what u studied, processing multiple inputs, manuals etc...

I cannot see how a human is faster at recovery from an error than an artificial intelligence that can get input in milliseconds and respond to all all the necessary components and controls without needing hands and fingers.

If you have a general AI that has an abstract view of the world like a pilot would have, then by all means. But we do not have anything close to a general AI.

VaxIFs6.gif
 
The day that AI figures out that the best course of action is to land itself on a river or to use asymmetric thrust for directional control is the day I will step onto a pilotless aircraft.

Humans will never be faster than AI but AI can't make s**t up to save itself.

What part of "intelligence" in "artificial intelligence" do you fail to grasp?
 
Some of you guys seem to struggle to get your heads around reality - pooh slinging makes no difference. We know how you feel about AI, fair, but let's be completely real here. Unless someone can actually make some solid points addressing the concerns about AI not being able to think outside the box or cope in the third world or outside of SOP or whenever there's glitches, then your arguments have no real substance. Nobody is saying we'll never ever see autonomous airliners - we're talking things in their current state.
 
Some of you guys seem to struggle to get your heads around reality - pooh slinging makes no difference. We know how you feel about AI, fair, but let's be completely real here. Unless someone can actually make some solid points addressing the concerns about AI not being able to think outside the box or cope in the third world or outside of SOP or whenever there's glitches, then your arguments have no real substance. Nobody is saying we'll never ever see autonomous airliners - we're talking things in their current state.

The original topic comment was ...

"Boeing has started researching commercial-passenger jets...."

So we're talking about the future, since that comment says "has started" not "currently".
 
The original topic comment was ...

"Boeing has started researching commercial-passenger jets...."

So we're talking about the future, since that comment says "has started" not "currently".

Let me help you with the egg on your face...

>Read the article

Nobody in this thread has said that AI will never ever be good enough to pilot autonomous airliners.

Debating the far future would be like you and me arguing over whether mankind will ever have a machine that could turn the moon into a solid block of cheese. You can't prove that there will never be such a machine, but the burden of proof wouldn't be on you in that case.
 
Just when you think the egg-heads couldn't be more stupid ...

I'm not anti-technology, but I am decidedly anti- stupid technology.
 
*GOOD...MOR...NING, LADIES...AND...GENTLE...MEN. WE...HAVE BEGUN OUR DESCENT...INTO...CHICAGO...O'HARE...INTER...NATIONAL...AIR...POR-OR-OR-OR-OR-OR-OR-OR-OR-OR-OR-OR-OR-OR-OR-OR-OR...*

kitt.jpg
 
Top
Sign up to the MyBroadband newsletter
X