Human vs. AI
 

I had an argument with my Aviation Logistics professor yesterday. He believes that within his lifetime (he's around 40-50) he will see pilots replaced by computers, probably some form of AI. Personally, I seriously doubt that we will ever be replaced. It would take some super form of a Bio-Neuro-Mechanical brain to have the decision making abilities of a human.

His argument was that if there are trains and trucks that are run entirely by computers, doing the same jobs that humans did with a higher safety performance, why can't the technology be applied to airplanes?

I countered with: Computers will probably be able to take-off, follow routes and land as well or better than a human. But computers base every action from a response. If events take place that confuse the computer, it will be unable to get past the diagnosis of the problem and won't reach the point of making a decision. Where as, a person may never be able to diagnose the problem, but they will be able to say, "Oh S#$%, I don't know what happend but I do know that I've go to land NOW!"

Personally, I'll never trust computers more than humans. Simply because I know that the human wants to survive just as much as I do in an emergency.

The only completely automatic trains I know of are at the airports for shuttle duty between the main terminal and the outlets. There aren't any freight/pax trucks/cars/planes that are full auto etc.... There has to be human over-sight, for the reasons you stated above. Could it be single-pilot large transport aircraft, I believe so, but it'll be awhile. Plus the general public wouldn't stand for less than 2 pilots up front for a long long time.

While the technology exists to operate unmanned transport category aircraft, it is very unlikely that this concept will be used in the civilian sector anytime soon. The main barriers will be the general public’s perception of safety and the exposure to liability that such an operation will create.

As you pointed out earlier, a computer can be programed to respond to certain inputs but in the framework of the dynamics of flying, judgement cannot be programed. The human element is still necessary to safely operate airliners.

Flying unmanned jetliners will require a totally redesigned air traffic system, including navigation, radar, and physical facilities. Although possible, it is highly improbable that unmanned passenger airplanes will be part of aviation’s architecture in any of our lifetimes.

P

I give it at least 100 years. There are many problems...

1. Judgement: No automated systems (even the best super computers) have the ability to respond to circumstances outside those which their designers anticipated. This is why the only automated ground vehicles operate in extremely controlled environments where there are going to zero outside factors (airport trains, mining trucks). This is the real kicker...it's not a question of applying current technology, this technology does not even exist yet. Sioux City type systems failures are very rare, but weather requires a lot of judgement on a daily basis.

2. Reliability: Current UAV systems have a non-combat loss rate of 1-2%. Airlines are more like 0.00001%. Any engineer can tell you that the first 95% is the easy part...after that it gets exponentially harder to ensure near-foolproof reliability. Airlines have that because the pilot can always hand-fly the airplane (and you have a spare pilot).

3. ATC: The entire ATC system would have to be converted to a system capable of managing and interacting with automated aircraft. In addition to technical and cost hurdles, the government red-tape will add at least 20-30 years to the whole process.

4. Security: An automated airliner would need a back-up ground-control link...if the bad guys jam it or hack it, all kinds of unpleasant things could happen (imagine 5,000 pre-launched cruise missiles awaiting targeting instructions :eek::eek::eek:).

5. Cost: The cost and effort to develop and certify this sort of system would be comparable to the manhattan project...no single airline is going to be able to afford it, so who's going to pay? Managers would rather just spend $1 Billion on pilots this year than spend $10 Billion on a program that might save the company money in 30 years. I'm not sure the government would see a benefit either (spending billions to eliminate 80,000 pilot jobs???).

It is remotely possible that you would see single-pilot cargo airplanes in our working lifetime (at the end). If you go to work for FDX/UPS it is vaguely possible that your last few years as a captain might be lonely ones.

We can help delay the process by demanding contract language that the airline neither fund, encourage, or participate in R&D programs which would target pilot jobs. The managers will cheerfully agree because it's a long term issue and won't affect THEIR stock options in the next few years (after which they'll be long gone).

The thought of a fully automated airliner should not intimidate anyone. Just like it was already stated, passengers would not stand to go on an airplane with no pilot. I personally would not unless it was the only way to get from point A to point B. Also stated above, the ATC system would have to get yet another makeover when the debate to fund it for just modernization is in full swing. I believe that too much will go into fully automated airliners and we'll not see them for a very very very long time.

It can, but what he isn't looking at is that trains and trucks have fewer variables when it comes to normal or emergency operations. If I remember correctly, I remember hearing somewhere that the military has a few UAV's that are pilot-less and run completely by computer programed parameters. It takes weeks for a whole team of programmers to program the mission and contingencies... and this is just for one flight. Imagine the man power needed to program the thousands of flights in each aircraft type/operation each day not to mention the extreme infrastructure changes that would be necessary to facilitate mass UAV operation. On top of that, you would certainly have mass resistance from the traveling public who wont trust these new machines without years of proven safety statistics. Not likely within anyone's lifetime that is alive today.

Thank ya'll very much. I'll take these to the professor and maybe he'll stop challenging me.