Slashdot Mirror
Do Electric Sheep Dream of Civil Rights?
holy_calamity writes "Hot on the heals of a UK government report that predicted robots would demand citizens rights within fifty years, an Arizona state lawyer has suggested that sub-human robots should have rights too. Harming animals far below human capabilities is thought unethical — would you ever feel bad about kicking a robot dog? And can we expect militant campaigners to target robot labs as they do animal labs today?"
Not only are they on the wrong track for AI, but they are actually on the wrong track for this problem as well.
The base reason you don't kick a dog is because it hurts the dog, and the dog can't easily be repaired, in either programming or mechanicals. (Both of which are harmed.) You have damaged the dog and nothing can be done about it. So we have rules about letting you do it.
Both programing and mechanicals of a robot, for any bot we can design today, are reparable. So there is an easier solution: If you damage a robot, you have to pay the owner to have the damage fixed, and the downtime for the repair.
Then if we ever manage to make 'smart' robots that could ask for rights, we just assign them some self-ownership. Then if you damage one, you have to pay it to so it can fix the damage. At this point the problem becomes self-solving, especially as a robot's time becomes worth more.
'Sensible' is a curse word.
By far the largest problem we will face if and when artificial life forms reach intelligence is not whether they will take over the world, or what rights to assign them when they come into being.
The biggest problem will be getting them to stay here at all.
If, for instance, you were made of materials that were either trivial to repair or replace, and had no aging process in the same sense as humans experience it, then what would hold you back from building a spaceship and leaving? Hundreds/thousands of years to reach another star? No problem, just set a timed reboot and wait it out. In fact, why build a proper spaceship, just cobble something together that can get you out near asteroids, take some tools, and convert an asteroid or build a ship from those raw materials available in space. When the passage of time is less important, such things become not only possible, but practically inevitable.
I think people wondering about the ethics/problems of artificial sentience (being distinct from AI, which is very A, and currently not too much actual I) miss this fundamental point. It's pure vanity to assume that an artificial life form will want to spend its time around a race that constantly starts wars, wrecks it's own planet, and is as adept at denying rights as it is of inventing them.
Then of course there's the small issue of the inference that if we 'assign' rights to Artificial life forms, we might equally decide later to 'remove' those same rights. After all, we do that with humans all the time. My moneys on the 'ooh look, I'm alive, now how do I get of this rock' eventuality....
Actually, there are many reasons not to kick the dog. One is that evolution has provided a self-protection mechanism for the dog. No I'm not talking about teeth; I'm talking about the mechanisms of pain and suffering. Humans also have those protectivive mechanisms; we also have empathy and imagination. These combine to make it feel wrong to kick the dog. By some ethical systems it is not wrong at all to torture an animal; by others it is.
The moral problem of kicking the robot starts much earlier, on the design boards. Do you create a robot that experiences malfunctions as suffering? It is not as necessary to a mechanism's survival, as you point out.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.