when it comes to human lives, there will always be human back-up. There was an article regarding robot and military I read a few months ago. Not only that, human must be in the loop, not on the loop. It's the term that military use.

as far as everything goes, it boils down to cash factor. if it's cheaper to use human, it will be human. if it's cheaper to use machines, it will be machines. what stopping from machines to run mc'd is human to do all the errand is cheaper than to use machines.

assembly lines is different as it usually tailored made for specific, and the whole resulting product is still went through quality testing.

and yes, robot like Predator (the autonomous jet thingy) does reduce the number of pilot needed. It actually cause the death of Comanche helicopter project.

"[font=sans-serif][size=2]The RAH-66 program was canceled in 2004 before it was fielded."[/font]
[font=sans-serif][size=2]"The Army also planned to use Comanche program funds to speed up unmanned aerial vehicles (UAVs) development.[sup][11][/sup] "[/font]
[font=sans-serif][size=2]
[/font]
[font="sans-serif"]Anyway, to be on topic, autonomous "robot" and self serve machines will reduce number of workers.[/font]

[quote name='slayemin' timestamp='1304035557' post='4804209']
However, you'll never be able to completely replace a human workforce (who will maintain & run the systems?).

If a computer can detect a faulty component, and a robot can build a another robot on an assembly line, then what makes you think that it is impossible, or even that hard, to build robots designed to service other robots?
[/quote]

Ok, but who will maintain those? O_o

More of the same maintenance bots?

It's entirely possible to fly an airplane from Point A to Point B on autopilot. If that's possible, why do they still put pilots in the the cockpit? I think the answer to my question is the same as the answer to your question.

See a list of autonomous UAVs. Or Russia's version of the space shuttle program.

We had the technology to fly heavier than air vehicles in 1903, so why didn't we have commercial commuter planes the next year? Because the technology was new and hadn't been proven yet. Commercial pilots are proven and have been shown to work, full computer controlled systems work, but are not yet 'proven'.

See a list of autonomous UAVs. Or Russia's version of the space shuttle program.

We had the technology to fly heavier than air vehicles in 1903, so why didn't we have commercial commuter planes the next year? Because the technology was new and hadn't been proven yet. Commercial pilots are proven and have been shown to work, full computer controlled systems work, but are not yet 'proven'.

Autonomous UAVs are extremely restricted in their usability and perform far worse than a human controller. That's why for most purposes they have the UAVs do the easy stuff autonomously and have humans take over for the tricky parts like take off, landing, or the actual mission objectives.

[quote name='Luckless' timestamp='1304273710' post='4805132']
See a list of autonomous UAVs. Or Russia's version of the space shuttle program.

We had the technology to fly heavier than air vehicles in 1903, so why didn't we have commercial commuter planes the next year? Because the technology was new and hadn't been proven yet. Commercial pilots are proven and have been shown to work, full computer controlled systems work, but are not yet 'proven'.

Autonomous UAVs are extremely restricted in their usability and perform far worse than a human controller. That's why for most purposes they have the UAVs do the easy stuff autonomously and have humans take over for the tricky parts like take off, landing, or the actual mission objectives.
[/quote]

As someone who has worked on the hobby side of autonomous UAVs, and seen what a few friends can do in a garage, I have to put a [citation needed] tag here.

The only thing a computer does worse than a human pilot is visual processing, but that is a general problem with robotics/AI these days. A human can easily look at an image and say "Yes, that's a car, and that's a tree, and they are all things I likely should avoid flying into". Where as a computer can easily look at its input data and say "Yep, that is stuff I shouldn't fly into". There are actually several aspects of flight where AI is far superior to human pilots already. (They don't get tried, they don't get distracted, they don't suffer from dial misinterpretation such as reading something as ascending when you are actually descending.)

We have been working on the problem of autonomous flight for less than 50 years. Powered flight was something that took us centuries to get off the ground. A computer isn't ready to plunk down in place of a human in all cases today, but we're really not as far away from it as a lot of people want to believe. (We're also not nearly as close as some people expect us are. 25-200 years seems a reasonable expected time frame, as opposed to next week vs never.)

I would welcome more automation for mundane tasks, but human experts are better than AI at dealing with unknown and unforeseen circumstances. So I would certainly welcome automated cars (autoautomobiles?), but there will be rare situations when it's better to take over manually, for example because you have to get off the road for a while, or come into some other unusual situation that the AI was not prepared for.

i guess panhandlers are obsolete now :-)

http://news.cnet.com/8301-17938_105-20018512-1.html
http://news.cnet.com/humans-cant-resist-giving-to-panhandling-robot/8301-17938_105-20057893-1.html
http://timpryde.com/blog/

when it comes to human lives, there will always be human back-up. There was an article regarding robot and military I read a few months ago. Not only that, human must be in the loop, not on the loop. It's the term that military use.

In "Wired for War" by PW Singer revealed that UAVs do, in fact, kill people automatically and have been from the beginning. The whole "in the loop" thing is propaganda.

It's a boring read, since Signer interviews several people about the same things, and they give pretty much the same answers over and over and over again without a unifying narrative. However, Part I is worth it and if you're interested in robotic warfare, it's worth a library check out.

i don't really know because i don't remember much. but i think due to the fact it kill human from the start is the reason they was planning to change it from "on the loop" (as in human just watch) to "in the loop" (they can acquire target, but a human must give them the OK-go order). The reason was due to some weapon testing in Africa where the automated weapon just shoot around at people, killing and maiming many, and another incident where a tank robot automatically target friendly just moment after deployed. the reason why that robot didn't shoot was still a mystery. i don't know if it just a propaganda, but i don't think military want autonomous robot that kill friendly (or civilian).