Aviation Question. Would you trust your life to a computer?

[Rei40c]

So browsing around there appears to be a push for a truly autonomous airliners. The computer tech to do so already exists. Some are saying we will see the first of these within our lifetimes. Essentially a drone. A drone which can make correct decisions hundreds of times faster than any human. Is immune to fatigue, stress, or having too much to drink the night before ect.

Would you trust it with your life? Your families lives?

I've thought about this and personally I still say no. While I understand it's superior in every way compared to a human pilot, I still want a human in the loop. What do you think?

 

[Jeepnik]

There is no way a computer could do a water landing on the Hudson. That takes the kind of judgment no machine is yet capable of making.

 

[Rei40c]

Don't all air travelers already trust their lives to computers (in the cockpit, in the tower, etc.)??

Yes, I'm thinking more about the idea of a airliner with no human control though. While I hear pilots are mainly system managers more and more, I wonder if people would accept the idea of them not being present at all.

 

[Rei40c]

There is no way a computer could do a water landing on the Hudson. That takes the kind of judgment no machine is yet capable of making.

Agreed

 

[CGDustDevil]

Some airliners (Airbus for instance) already have the auto-land capability which is used, although not often. If you fly a lot, you may have already survived an auto-landing!

 

[Colonialgirl]

"Welcome Aboard"; As we taxi away from the terminal for our take-off, We ARE PROUD to announce that there are NO pilots in the cockpit. Before you panic please know this IS the FIRST Flight completely controlled by a computer and nothing can go wrong, go wrong, go wrong, go wrong,go wrong.

:shock: :shock: :shock: :shock: :shock: :shock: :shock: :shock: :shock: :shock: :shock:

 

[toysoldier]

My son had lots of hands-on experience with computer hardware before he earned his PhD in computer science. He teaches at a private college---using written notes and a chalkboard. Why? "Because it always works."

 

[Snake45]

I've never owned or used anything that used electricity that didn't fail every so often--usually at the worst possible moment.

This is why I will never have an electronic "smart gun," and why I won't count on lasers or lights to substitute for good old hard-won shooting skill.

 

[JFB]

are you aware there are currently cars driving their self?
http://www.reuters.com/article/us-tesla-autopilot-idUSKCN1062CT?utm_campaign=trueAnthem:+Trending+Content&utm_content=5797c97504d30130f36c4526&utm_medium=trueAnthem&utm_source=twitter

 

[Fox Mike]

I can almost assure you if that happens Murphy will receive his pilot's license. The big difference between a drone and an airliner is if the something on the drone goes South you only lose the drone. In a airliner there are real people to consider. I believe I would like to have someone there with 'hands on' ability is something breaks.

 

[427mach1]

The computer control is only as good as the flight control programs and parameters, which are input by humans. The flight control laws are derived from wind tunnel data, and the wind tunnel data covers a finite flight envelope. If the airplane goes outside the envelope into areas where wind tunnel data were not collected, the flight control might not know what to do to recover. There was a recent incident where a military transport got outside the normal flight attitudes but the pilot was able to recover. The airframe suffered structural damage and was no longer flight worthy. Had there not been pilots onboard.........

 

[737tdi]

IMO it will be decades if ever that an airliner will fly without pilots. Yes, they can fly themselves and 90% of the time they do. But in that 90% there are many times that a problem will occur and the auto flight will disengage. Yes you need a pilot then. A lot of airlines use auto land but I can say mine doesn't. Company policy is the auto pilot will be disengaged and approach stabilized before 1000' AGL. Technology is a great thing but it is always subject to failure. We see it every day. Aircraft are reliable but they have a fairly rough life, especially for operators like SWA who fly a lot of short haul. All of those cycles take their toll on everything.

Karl

 

[Rei40c]

One of the things that got me thinking along these lines is a new AI fighter pilot the Air Force has been developing called "Alpha". Recently they put it up against a Airforce pilot who from what I gather is regarded as not only the best we have had recently. But perhaps the best fighter pilot in decades. The now retired U.S. Air Force Colonel Gene "Geno" Lee. He's been active in USAF testing of AI going back since the 1980's. He could always beat it, until recently. They ran hundreds of simulated dogfights and Colonel Lee was defeated every single time without scoring a single simulated kill. He stated it felt as if the was as if the AI was "reading my mind" and knew the exact moment to transition between offense and defense. Finally in frustration they crippled the AI to a level of %50 of it's computational power and tried again. Same result. Hundreds of simulated dogfights, Col. Lee was shot down ever single time, while scoring 0 kills.

So I think we may have reached an important point in which AI has advanced to a point where people are asking why to bother with a human at all.

There's many easy to find links on the story here is one. http://www.popsci.com/ai-pilot-beats-air-combat-expert-in-dogfight

 

[RSIno1]

A friend flew L-1011s it's auto land worked so well you had to smack the yoke after it landed or the brakes would not actuate. Even if they come up with a self fly system they will still have a human presence in the cockpit for the first 20-30 years just to make you feel safe.

 

[wizofwas]

It's not really a mater of trusting the computer, it more of trusting the computer programmer. And if it can be hacked, it will be hacked, if someone wants to do harm to passengers and/or cargo onboard.

 

[blume357]

I remember growing up one of the most common phrases was:

"Computers don't make mistakes"

when was the last time you heard that? I think Bill Gates and Steve Jobs proved that wrong....

 

[wizofwas]

LOL, computers don't make mistakes, they can only do what they're told to do. If that is wrong, then it's not the computers fault, it's the programmers.

 

[Dan in MI]

I used an early prototype version of an automatic braking system. We set up a cardboard box as a pedestrian to test it for fun, nothing official just a "let's see if it works" thing as this wasn't something my group normally works on. I come rolling up at about 5 mph and put my foot under the seat to prevent me from instinctively hitting the brake. The car stopped just fine and the box lived. As I laughed and said This is cool, the car started moving again and rolled over the box. Yeah, it stopped the car, but I didn't put my foot on the brake after it stopped. If goofy things don't get caught in testing they will hit production.

Mark me as not a fan of automated flying/driving. A train I can trust somewhat as it is on tracks, but I'm still not 100% sold. If I want a thrill ride I'll go to Cedar Point and trust their computers. At least then I signed up for a thrill.

 

[RonEgg]

So much of what is wrong with aviation today is the computer and automation. It flies the aircraft better and more efficient that a man can and bean counters then mandate "going auto" as soon after takeoff as practicable. The result? Lazy pilots, not paying attention and picking up subtle changes during the flight. When your talent is needed, it is rusty or non existent.

 

[NixieTube]

If we haven't reached the "inflection point" we will soon. The big question everyone in society is going to have to ask themselves was first discussed back in the 1960s at big universities: "What are people going to do?" regarding advanced AI of the kind we're going to see. We've already got robots that can build a house in 1/5th the time of a normal group of builders. That's going to go to 1/20th the time. At that point, what do you do with the guys in the pickup trucks?

It's very apt that you ask: "Why bother with a human at all?"

People are going to have to decide - within the next 10 years - how they are going to deal with machines that replace them on a massive scale. It's going to cause really serious problems in society, not just in the pilot cockpit. Almost every single thing our society is based on from prenatal care to continuing education, medical care to undertaking, is grounded in the reality of the human lifespan and what human beings can do. Our limitations - and indeed, the phases of the human lifespan - provide the framework for how our society is structured. Once we have machines that surpass us in almost every field of endeavor, the answer to the question "Why bother with a human at all?" is a very dark one. It literally disrupts the world from top to bottom. There are a lot of really selfish people who look at that and enjoy the prospect, but I don't.

If you think about trying to accomplish something as a person, it becomes even more acute. Why bother studying when there's an AI system that already has the knowledge it's going to take you 2,000 hours or 3,000 hours to accumulate?

At that point, we're not just looking at these machines making humans more productive - we're looking at them making humans completely obsolete. I don't have the credentials to make a credible, media-worthy statement about it, but there are a few people who are seeing a little farther than ordinary and they're correct.

Obsolete humans have nothing to do except drugs. I do not buy for one instant the idea that the massive increases in unemployment will be offset by an increase in demand because of efficiency. That's a pipedream. I expect a lot more people on drugs and the public dole by 2040, and it's not going to be a nice world, except for the phenomenally wealthy. We have about 10 years to make the hard choices that are going to be necessary.

 

[427mach1]

Here's what can happen when the autopilot has more authority than the human pilot. This is the one thing I really don't like about Airbus aircraft.


https://www.youtube.com/watch?v=-kHa3WNerjU

 

[Snake45]

One of the things that got me thinking along these lines is a new AI fighter pilot the Air Force has been developing called "Alpha". Recently they put it up against a Airforce pilot who from what I gather is regarded as not only the best we have had recently. But perhaps the best fighter pilot in decades. The now retired U.S. Air Force Colonel Gene "Geno" Lee. He's been active in USAF testing of AI going back since the 1980's. He could always beat it, until recently. They ran hundreds of simulated dogfights and Colonel Lee was defeated every single time without scoring a single simulated kill. He stated it felt as if the was as if the AI was "reading my mind" and knew the exact moment to transition between offense and defense. Finally in frustration they crippled the AI to a level of %50 of it's computational power and tried again. Same result. Hundreds of simulated dogfights, Col. Lee was shot down ever single time, while scoring 0 kills.

So I think we may have reached an important point in which AI has advanced to a point where people are asking why to bother with a human at all.

There's many easy to find links on the story here is one. http://www.popsci.com/ai-pilot-beats-air-combat-expert-in-dogfight

With all due respect, that is quite a different scenario from safely delivering a couple hundred passengers--and doing it fitting into an intricate schedule, and doing it thousands of times a day, worldwide.

I expect that in our lifetimes we'll see pilotless fighter and attack aircraft, and we pretty much already have pilotless recon birds. Pilotless airliners? Not seeing it.

 

[Rei40c]

If we haven't reached the "inflection point" we will soon. The big question everyone in society is going to have to ask themselves was first discussed back in the 1960s at big universities: "What are people going to do?" regarding advanced AI of the kind we're going to see. We've already got robots that can build a house in 1/5th the time of a normal group of builders. That's going to go to 1/20th the time. At that point, what do you do with the guys in the pickup trucks?

It's very apt that you ask: "Why bother with a human at all?"

People are going to have to decide - within the next 10 years - how they are going to deal with machines that replace them on a massive scale. It's going to cause really serious problems in society, not just in the pilot cockpit. Almost every single thing our society is based on from prenatal care to continuing education, medical care to undertaking, is grounded in the reality of the human lifespan and what human beings can do. Our limitations - and indeed, the phases of the human lifespan - provide the framework for how our society is structured. Once we have machines that surpass us in almost every field of endeavor, the answer to the question "Why bother with a human at all?" is a very dark one. It literally disrupts the world from top to bottom. There are a lot of really selfish people who look at that and enjoy the prospect, but I don't.

If you think about trying to accomplish something as a person, it becomes even more acute. Why bother studying when there's an AI system that already has the knowledge it's going to take you 2,000 hours or 3,000 hours to accumulate?

At that point, we're not just looking at these machines making humans more productive - we're looking at them making humans completely obsolete. I don't have the credentials to make a credible, media-worthy statement about it, but there are a few people who are seeing a little farther than ordinary and they're correct.

Obsolete humans have nothing to do except drugs. I do not buy for one instant the idea that the massive increases in unemployment will be offset by an increase in demand because of efficiency. That's a pipedream. I expect a lot more people on drugs and the public dole by 2040, and it's not going to be a nice world, except for the phenomenally wealthy. We have about 10 years to make the hard choices that are going to be necessary.

Well said. While I have a hard time embracing the new role of computers and the robots they will command, maybe this is not a question for our generation but the one just after us. Our children will likely be the ones that really have to grapple with these tough issues. I believe they will also be more accepting of the greater control machines have over our lives, and that we probably never will accept it.

I've seen this myself watching my niece who had a computer tablet put into her hands at age 7 and actually learned to use it fairly well. That person will grow up to be more accepting of some of these ideas than many of us here who grew up with different backgrounds.

 

[Rei40c]

So much of what is wrong with aviation today is the computer and automation. It flies the aircraft better and more efficient that a man can and bean counters then mandate "going auto" as soon after takeoff as practicable. The result? Lazy pilots, not paying attention and picking up subtle changes during the flight. When your talent is needed, it is rusty or non existent.

To bring up the Hudson river landing again, I remember many pilots openly admitting that if the public thinks this is the general skill level of most commercial pilots they are mistaken. Sully was an oddball (in a good way). He started flying as a teenager and continued through his entire life.

As you said from what I gather, chances are excellent you will not have a "Sully" as a pilot for your next flight unfortunately And yes I think that problem will get worse over time.

 

[RonEgg]

With all those Chinese chips, what could possibly go wrong???

 

[gramps]

"Sully" was a competent sail plane pilot also. I for one will never fly or ride in an automatic plane or car. Computers fail way to often for my comfort.
gramps

 

[NixieTube]

Well said. While I have a hard time embracing the new role of computers and the robots they will command, maybe this is not a question for our generation but the one just after us. Our children will likely be the ones that really have to grapple with these tough issues. I believe they will also be more accepting of the greater control machines have over our lives, and that we probably never will accept it.

I've seen this myself watching my niece who had a computer tablet put into her hands at age 7 and actually learned to use it fairly well. That person will grow up to be more accepting of some of these ideas than many of us here who grew up with different backgrounds.

My Dad was an IBM engineer and I was the first person to carry a laptop computer to school while I was in high school. I was surrounded by computers as a kid, ranging from TRS-80 Model 1s to the first Atari systems to IBM 4380 series mainframes. I'm very far from a Luddite - I'm talking with you over a TCP/IP connection!

But there are things we need to grasp - for the benefit of what we call humanity - that we should not reasonably do. In the late forties and early fifties, for example, there were plans to build aircraft powered by nuclear reactors. In the planes, like aircraft carriers. They probably would have worked! But the consequences of having even one of them shot down and blown to bits over thousands of square miles were so awful that we (can you believe it?) never built them. https://en.wikipedia.org/wiki/Nuclear-powered_aircraft

More to the point, I know now that there's a big difference between making people's lives less toilsome, getting rid of the soul-crushing stuff, and actually crushing their souls by replacing them without any idea of what comes next. People need to work - they're defined as much by their careers and what they do during the day as they are by their own personalities - and that is what keeps them moving and keeps them sane (largely speaking.) This is true for the bricklayer and the brain surgeon, for the professional philanthropist and the person who picks pistachios. It's even true for people like Woody Allen, who probably wouldn't be willing to live any longer without his work schedule.

Our idea of a society we want to live in revolves around people being able to apply themselves and realize their potential (at least in the abstract that's the goal.) If you take that away from people - millions at a time, in a very short time span, I don't know what you really have. Our society is founded on "Life, Liberty and the Pursuit of Happiness." What does that mean to an AI-guided robot that takes 20 people's careers from them?

Past technological changes are fundamentally different from the ones we're going to see in the next 20 years. Charlie Chaplain's work was a very long time ago and he didn't and couldn't understand what unprecedented changes would be wrought by AI and robotics in the next 20 years.

I respect people who do their jobs and aspire to do them better, whether they're painting houses or painting works of art. I'm very wary of the "Things to Come." We've haven't taken the time to think through this juiced-up, hypercharged by finance capitalism (and I'm a capitalist in a big way) effort to do this As Fast As Possible.

https://www.youtube.com/watch?v=JOK0v2QYy9Y

[Addendum: I remember reading Ray Kurzweil when I was a kid and while his musical products have always been astounding I've also thought, throughout my entire life, that some of his other ideas were borderline sociopathic. http://kurzweil.com/ He wants to upload his consciousness into a computer of arbitrarily advanced design. I don't think he should try.]

 

[Scout943]

Today's autos (they don't build cars anymore ) are loaded with computers nothing ever goes wrong with them right ? ? :shock: :shock: :shock: :roll: :roll: :roll:

 

[NixieTube]

Back to the point of this thread.

My answer is: NO.

I won't get on one of them. I can handle human error, if it happens it happens. It's very rare, and we do a good job of managing it.

 

[exavid]

I'd rather trust an autopilot over a Korean flight crew, especially landing at SFO.