http://en.wikipedia.org/wiki/Autoland
http://en.wikipedia.org/wiki/V-1_flying_bomb

autonomous modes for some time now. However, the ability to do something as difficult as land on an aircraft carrier is definitely a step toward total autonomous operations--from a technical standpoint, at least. The political and ethical implications are something else entirely!

programmers screw up does not give me much faith in such devices functioning as intended - even assuming the intents were acceptable.

Give the use of a drone to kill an unindicted, uncharged American citizen and the provisions of the latest NDAA act, maybe it's time to develop a cheap, civilian-use anti-drone technology for self-defense.

It is worth remembering that the Founding Fathers were all traitors.

I believe the Air Force's UCAV is designed to fly in formation, and to dogfight autonomously if needed, with one human pilot in an F-23 bossing up to 6 of the high-performance drones.

However, as a couple of my ex-military friends have pointed out recently on Twitter, the one universal red line so far is that a human being has to be involved in any "weapons free" decision loop. None of them felt that was a line the military would cross anytime soon.

is being ported to both commercial airlines and the auto industry. Autonomous control is becoming a huge potential add on.

Airline flight is autonomous except take off and landing. These technologies add autonomous takeoff and landing.

In autos, there is the potential to bring traffic fatalities to under 5% of what they are currently. That is the huge payoff in autonomous auto transport. It's fascinating to watch the automobile in society. There is no other product that kills more people per day which goes without any backlash or perception of fear. Can you imagine if a plane crashed every other day. Would you fly?

It reminds me of the study done about birds and wind power. It was found that wind turbines were killing 150,000 birds a year in California and presented as a reason for ending these projects. But then someone pointed out that cars are killing 150,000 birds a month and there is no reaction whatsoever. Cars explode in flaming balls of fire and it does not make the news, but if the battery catches fire! International incident!!

Ideally, we'd have a world where people "get it" and this use of technology to destroy each other and/or our property would be out of the question.

But, we don't live in that world. So the next best thing, I guess, is to minimize the human casualties (at least on our side).

While the ethical question is important, I'd like to point out a few things that are already big problems in the current (non-robotic killing machines) era.

First is the responsibility question.

Right off the bat I want to point out that humans are NOT better decision makers than the robots.

Humans made the decision to wipe out My Lai in Vietnam. Humans made the decisions to go on unauthorized revenge-killing sprees in Afghanistan. Humans routinely make not just morally wrong, but also strategically bad decisions about detaining and torturing innocent people. Guantanamo is a human-decision-caused institution, and so are all the CIA black prison sites where detainees are sent and tortured without any observance of their human rights.

Thus, the argument that humans are better decision makers than a properly programmed robot is full of holes you could drive an aircraft carrier through.

Beyond that, when it comes to operations demanding mechanical exactness, ability to focus on details, calculate parameters, and fine tune activities such as takeoff or landing, the robots will in all likelihood, outperform even the best human pilots.

Perhaps there will be a few botches while they perfect the algorithms, but I'd say within 5 years, and for sure by 10, the idea of having a human control a takeoff or landing will seem very stupid, unless there is a total onboard failure of the automation system (such as if a drone gets hit by enemy fire).

Next is culpability. Already it is VERY difficult to pin the tail on the donkey when unauthorized killings happen on the military's watch. The SOP is to throw the lowest ranking meat possible to the prosecution dogs, leaving the real decision makers in high places unscathed, despite repeat offenses. The Pat Tillman affair, Abu Ghraib, etc are all examples of this sorry "I didn't do it" culture we have created.

So, already, we do, at best, a VERY mediocre job of correctly assigning responsibility in cases of military ventures gone horribly wrong.

In spite of all these points, it IS a vitally important issue, so I am glad they are working on it.

But in reality, I'm not expecting too much different from the responsibility-evading mush-mouthiness and failed, politically-tampered-with "investigations," that we already see in high profile cases where the wrong people end up dying for no good reason.

If you can't hold the robot responsible, then it falls back on the humans, who, of course, will pass the hot potato around as fast as they can shove it off on somebody else.

That, I am afraid, is something the technological advancement is not going to change.