Robosoldiers
This is pretty disturbing -- robots replacing soldiers in battle:
In the heat of battle, their minds clouded by fear, anger or vengefulness, even the best-trained soldiers can act in ways that violate the Geneva Conventions or battlefield rules of engagement. Now some researchers suggest that robots could do better.
"My research hypothesis is that intelligent robots can behave more ethically in the battlefield than humans currently can," said Ronald C. Arkin, a computer scientist at Georgia Tech, who is designing software for battlefield robots under contract with the Army.
The article goes on to describe research being conducted by Arkin and others on the potential use of autonomous robots in battle. It's important to note that this article is describing robots which make their own decisions in the heat of battle, not drones controlled by humans from a distance.
It's creepy, it's wrong-headed, and it's impossible to deploy without catastrophe.
Which is more horrifying: the surrender of human autonomy and agency to a machine, the threatened proliferation of that technology, or the implied trust by the scientists that such a technology could be made without critical and fatal flaws?
It would seem to me that a truly effective robotic soldier programmed with the rules of ethics and morality (as the article discusses) would require something approaching true artificial intelligence (if not outright artificial sentience), and that technology still lies very much in the realm of science fiction. No machine has ever successfully passed the Turing Test, let alone the more refined (and difficult) tests for artificial intelligence. Even those tests are based solely on language and communication and don't even begin to delve into the chaotic complexity necessary to respond to the nearly limitless stimuli of the physical world.
It beggars the imagination to suggest we're on the verge of creating a machine that could pass even a Turing Test set in the chaotic midst of a battlefield, much less the life-and-death decisions made by a soldier. I don't believe it for a minute.
There's no way that a binary-based computer (or the robot it controls) could reliably distinguish a genuine attacker from a child running toward it with a toy gun. There is no way a robosoldier could discern the gulf of difference between a hostile mob and a wedding party shooting rifles into the air in celebration (as is common in the Middle East). No robot programmed by humans could ever reliably and accurately decide on the fly whether a target engenders too much risk of collateral damage.
However, I do realize that half-baked science and ill-conceived technologies have a tendency to make their way onto the battlefield without adequate forethought; see the V-22 Osprey and SDI for starters. Given that track record of stupid technology deployment, the prospect of robosoldiers patrolling the streets of the next country we invade stands as all too stark and terrifying a possibility.
In my opinion, the scientists who believe this technology is possible without the risk of catastrophic consequences are either fooling themselves or angling for more Pentagon funding. Unfortunately, the politicians who fund the defense contracts in their districts likely won't bother to ask questions deep enough to prevent its deployment.