Which would be more dangerous?

[CID]

Rockie, your viewpoint seems more romantic than technical. "AI" isn't neccesary. In fact it would be less benificial. The whole point is to make the machine perform it's function without question. Once you put short term variability in the equation, the results are difficult to determine. Just like with human pilots.

The key is to have a very thorough algorithm that takes as many scenarios into consideration as possible.

[Legacy]

A bunch of you that think this will happen are under the assumption that a pilotless aircraft is going to cost the same as a piloted aircraft. Not even remotely close. I wouldn't be surprised to see it cost 100 million more. So not only do you have to pay a huge amount more for the aircraft you STILL have to pay for someone to fly it, or monitor it, remotely. So the benefit to cost, if any, might not even be there. No to mention who will be the first insurance company to insure and what will they charge until such a system is a proven one. The REAL savings in airliners is the fuel. That is what the future is going to dictate on the advancements in airliners.

[Rockie]

Once you put short term variability in the equation, the results are difficult to determine. Just like with human pilots.

I'm sure you noticed, but flying an airplane 24/7/365 in extremely busy airspace in all kinds of weather is chock full of short term variables.

The key is to have a very thorough algorithm that takes as many scenarios into consideration as possible.

Ah...if only it were as easy to do as it is to say....

[GyvAir]

Taking all the pilots out of the picture removes a very prolific source of short term variability.

Autopilots never fail to have their head in the game due to a cold, a bad burrito, divorce proceedings, dropping a device on the floor, one too many the night before, etc. etc.

No egos to help get into situations requiring drawing on extraordinary piloting skills to get out of.

A sudden snow squall blowing up on short final wouldn’t even trigger a need for a decision with no pilots involved.

Just to name a few..

[CID]

I'm sure you noticed, but flying an airplane 24/7/365 in extremely busy airspace in all kinds of weather is chock full of short term variables.

The short term variability I mentioned has nothing to do with combinations of weather and workload. It has to do with "decision making". Computers follow the directions given to them. Therefore, no short term variability like a conscience, or a headache or fatigue will affect the ability of the computer to follow the program. And with proper "watch dog" systems and redundancy, even failure shouldn't stop it.

Ah...if only it were as easy to do as it is to say....

Keeping an aircraft in stable flight, or landing or taking off is not a daunting task for computers and their programmers. Neither is avoiding CFIT. Or avoiding other traffic. Or following routing information. Or diverting around weather. Individually these are not insurmountable tasks. Together they can be difficult for human pilots but if there is something that computers are awesome at it's multi-tasking. With no additional stress to add short-term variability.

These are not just dreams. The technology is available today and pilots on the more advanced aircraft are now more system managers than airmen. In fact, for some of these airplanes like the 777, it's best to let the automation do it's thing. Without all that automation, it would be impossible to consistently safely operate such an aircraft on the routes it flies with just two cockpit crew members.

Remember, trans-oceanic flights used to have two pilots, a flight engineer, a navigator and an communications officer on board. Automation has reduced the crew to 2 even on aircraft that fly 4 times higher, 5 times faster, and with 10 times the passenger loads. And with a tiny fraction of the accident rate.

[xsbank]

They haven't figured out how to get the airplane to and from the gate. When that happens, pilots are toast.

Pilot error is the biggest cause of accidents. Lose the pilots, even safer.

[GyvAir]

They haven't figured out how to get the airplane to and from the gate.

Except for the difficulties encountered where snow and ice on the ramp are concerned, I suspect grafting some warehouse drone technology onto a fleet of tow motors would take care of that. And it would even make redundant the tug driver and wing walkers.

[complexintentions]

These are not just dreams. The technology is available today and pilots on the more advanced aircraft are now more system managers than airmen. In fact, for some of these airplanes like the 777, it's best to let the automation do it's thing. Without all that automation, it would be impossible to consistently safely operate such an aircraft on the routes it flies with just two cockpit crew members.

Remember, trans-oceanic flights used to have two pilots, a flight engineer, a navigator and an communications officer on board. Automation has reduced the crew to 2 even on aircraft that fly 4 times higher, 5 times faster, and with 10 times the passenger loads. And with a tiny fraction of the accident rate.

The extrapolation of technology is not a valid argument. As mentioned repeatedly, the technology has existed for decades for remotely piloted aircraft. Indeed, they have been used all along in niche applications. It turns out they're really suited for dropping bombs in Pakistan, for example. Yet, at the same time many things have changed in aviation, the basic design of airliners has remained the same. It seems there are some things that are optimum - or even "good enough" - and then it becomes more economic to simply refine elements of the design (i.e. aerodynamic advances, engine technology, construction materials, manufacturing processes, etc).

As a pilot of the B777, I can categorically state that it's not always best to "let the automation do its thing". The automation will quite happily fly the machine into the ground without human intervention, and will do it just as proficiently as it holds altitude in cruise. The automation is simply a set of systems to be managed, as you say. The only difference between that and "managing" flight controls directly is that the automation performs the repetitive, mundane tasks thus freeing up resources (time, mental acuity) for the humans to do what they still do better than algorithms - make judgements. How is this an argument for allowing the automation to be given complete control?

To suggest that having the humans on the ground doing essentially the same thing as they do in the aircraft, would be cheaper, is not really a defensible idea either.

As someone who works for a company that looks at virtually every possible way to maximize profit - and virtually limitless financial resources to realize their strategies - I can assure you that if pilotless airline flight was cheaper than piloted, it would already be happening. While it is perfectly plausible that pilotless airliners will one day appear, the arguments presented so far are not compelling. Some - like projecting the use of driverless shuttles between terminals at SFO onto passenger flight - are simplistic to the point of absurdity. People do realize there is still human monitoring of all driverless trains, right?

There would have to be a paradigm shift in more than one field simultaneously, i.e. the technology AND the economics AND the psychology AND the politics would all have to change, at the same time, and that is unusual.

I am not the buggy whip maker poo-pooing the horseless carriage. I'm saying it's not quite the incremental change some think it is, at all.

[Rockie]

The technology is available today and pilots on the more advanced aircraft are now more system managers than airmen.

Incorrect. Pilots are not only required to maintain an ability to use their hands and feet to fly the airplane, they are often required to actually do it when your vaunted technology is unable to keep up with the dynamics of flight which happens regularly. There are also electronic and paper checklists full of hundreds of procedures for the pilot to do when the technology breaks that contain lists of inoperative secondary equipment. Many times that equipment happens to be the autopilot among many more items requiring direct pilot intervention.

Pilot decision making is of course the other arena that technology will never be able to achieve short of artificial intelligence. And as I said before it would be a mistake to assume an AI won't make mistakes too.

Airplanes are not people movers affixed to rails unaffected by weather or other traffic. They operate in 3 dimensions at high speed and do not have the advantage of coming to an emergency stop when a simplistic electronic monitor detects a fault.

Excuse me for asking but are you a pilot, and if so what do you fly?

[photofly]

People do realize there is still human monitoring of all driverless trains, right?

The Docklands Light Railway in London runs automatically. There's a "train manager" on board each service, and that person is trained to take control if there's a requirement, but otherwise their main role is to patrol the train, check tickets and tell youths to take their feet off the seat cushions.

I can foresee a future where a flight is fully automated but the chief steward ("aircraft manager") is trained to operate it if human judgement is called for, and perhaps to take control for taxiing to the gate after the automatic landing.

Fairly soon getting a CPL may involve learning the right way to serve peanuts.

[YYZSaabGuy]

I can foresee a future where a flight is fully automated but the chief steward ("aircraft manager") is trained to operate it if human judgement is called for, and perhaps to take control for taxiing to the gate after the automatic landing.

So Henri wasn't that far off the mark with his "taxi specialist" references?

[CID]

Yes, the 777 will happily fly itself into the ground because it's not designed to be completely automated. However, the automated functions are done much more accurately and consistently than the human counterpart. I brought it up because the 777 is a representative milestone in the development of automation in large aircraft.

Yet, at the same time many things have changed in aviation, the basic design of airliners has remained the same.

Uh...no. If you mean there are usually two wings, you're right. Otherwise you can't tell me that a DC-3 is like a 787.

To suggest that having the humans on the ground doing essentially the same thing as they do in the aircraft, would be cheaper, is not really a defensible idea either.

I don't think I suggested that. But the space shuttle program proves it can be done. For the most part the computers (flawlessly) controlled all phases of flight with initiating commands from the ground. And they did all that with less than 1 Mb of digital memory.

Some - like projecting the use of driverless shuttles between terminals at SFO onto passenger flight - are simplistic to the point of absurdity.

Not absurd at ALL. Every complex systems have their predecessors. You can't launch into a fully automated airliner without proving various concepts on simpler levels. We already know that autopilots can fly airplanes more precisely and consistently than human pilots. The link between ATC and the autopilot is getting closer all the time. CPDLC might soon be pushing directly into the FMC.

I agree that there are plenty of things that need to align to permit pilot-less flight and I doubt I'll ever see it but there is a difference between possible and probable. Possible? Absolutely.

[xsbank]

I see a bank of simulators and a half dozen pilots on call in front of Netflix. "Captain Blackfly, report to Sim Bay 6. Flight 007 to London Heathrow is reporting an engine anomaly."

Saddle up, push the big red button and Captain Blackfly can fly the sick airplane into London, remotely.

[jump154]

Some - like projecting the use of driverless shuttles between terminals at SFO onto passenger flight - are simplistic to the point of absurdity.

Interesting discussion. For clarity, the point I was trying to make was not that the technology was possible but that the General Public is accepting of apparently fully-automated tranpostation systems - using the SFO shuttle and elevators as an example.

[Meatservo]

You guys talk as if simply being able to manipulate the controls is all there is to being an airman. Having flown over ten thousand hours on fully manual aeroplanes of all shapes and sizes before my type rating on a fully-automated one, I can tell you that there is no philosophical difference between interfacing with the aircraft through levers and pedals, and interfacing with it through the FMS and the MCP. Your brain is still there, in the front of the aircraft, hopefully guiding it, rather than being guided BY it, along its route. Before you enter a hold or an approach procedure, hopefully your mind has already travelled the route and decided what it wants. Hopefully your mind has interfaced with the automation so that what you plan to do and what it plans to do are one and the same. After that, the difference between hitting the "execute" button and watching your plan unfold and steering it along yourself should be no more or less dependent on your skill than taking a super-cub around the circuit. You all know how to fly a plane, don't you?

If they want to take the brain out of the front of the plane, fine. My job is secure anyway, and I say to those who want to phase the pilot out of the system, caveat emptor. In fifty years if I'm even still around, I will shell out extra for the manned (or womanned) option, and more or less gleefully watch this exact same argument go through its fiftieth iteration.

One thing they don't compile statistics on is how many times the automation would have ruined everybody's day if there wasn't a pilot to guide it. Because most rational pilots subscribe to the viewpoint I expressed above.

[Rockie]

Bottom line is if computers can ever replace the judgement and decision making of a human pilot, there will be no job they can't do. No more CEO's, mechanics, doctor's plumbers, teachers, investment bankers, soldiers, waiters...not even computer programmers. Humans won't be needed to do anything.

[Big Pistons Forever]

Piloting is still a blue collar job. There are many blue collar jobs that computers have taken over.

There is one bright side. If the pilots are gone management will no longer have a pool of employees they can randomly abuse knowing they will always take it........

[Meatservo]

I'm sure there's nothing intrinsic to a so-called "white collar" job that means it can't be automated, other than the obligatory faux-hawk hairdo, pointy shoes and faggy clothes.

Oh wait, there's the self-aggrandizement and self-promotion and "marginalization of your perceived inferiors"part. I guess computers haven't gotten sophisticated enough yet to pull that off.

Sorry if I'm swimming against the stream of your reverence for the upper class...

[Rockie]

Shoe makers used to be one of the most valued professionals in society and Doctors among the least. White collar and blue collar are merely a matter of perception - certainly not relevant to this discussion.

[Big Pistons Forever]

Pilots are machinery operators, that makes them blue collar. I don't mean that in a perjorative sense as it is honourable work that provides a benefit to society.

I hold both a professional pilots licence and a Bachelor of Science degree. The amount of work and the intellectual demands to get the university degree were an order of magnitude higher than my piloting qualifications.

I have worked on both sides, blue collar and white collar, and don't judge my personal worth, or anybody else's worth, solely by the qualifications they hold. But I also recognize that some occupations require more demanding and difficult to obtain qualifications than others. These distinction are just as evident in aviation, which is why the flight attendants in the airliner are not treated the same as the pilots.

However melodramatic portrayals of how piloting is somehow so much more special than a host of allied occupations serves no ones interest. Putting your head in the sand and saying that pilots can never be replaced requires piloting to be different than every other skilled trade, an observation that IMO, is simply delusional.