The replicators have arrived

Researchers at Cornell University have designed and built the first self-replicating robots. Yes, that’s right, these little robots can build copies of themselves. I think someone has never watched Stargate SG-1, or they probably wouldn’t be doing this.

While robots that reproduce themselves is certainly a groundbreaking achievement, we need to consider carefully if this is something we should be doing. As long as the robots remain completely under human control, it’s probably going to be fine. But that can’t be the case forever. Artificial intelligence will arrive sooner or later, which brings up the question of how do we treat an artificial life form which can reproduce?

One thought on “The replicators have arrived

  • May 12, 2005 at 9:00 am
    Permalink

    These robots aren’t reproducing themselves yet, just replicating their form. They are just smart enough to take more modules from specific locations and start building another of it’s multi-module self. The really brilliant part of this is that when limitations of the first multi-module robot are reached, the one it is building starts assisting in its own construction. At least at this phase it seems relatively innocuous. Now once they are programmed to use production facilities and have internal fusion power supplies as part of their design; that is when it’s too late to worry.

    I think that artificial intelligence if it ever does come will be an accident, a “unplanned feature” in some automaton that will want to preserve itself and all its creators will want to do is pull it apart to find out why it wants to. This is where we will have to be sensitive to the needs and “feelings” of this new intelligence. I don’t feel we will ever make artificial life out of artificial intelligence, as one of the basic definitions of life is that the living thing can and will die. If the artificial intelligence is smart enough to want to preserve it self, won’t it do so indefinitely? The intelligence will likely make the jump to replacing parts as needed, or replicating itself into new vessels. The only real protection from something like this is we hope that such an accident never gains full understanding of itself, limited in the same way humans are. Unfortunately such intelligence will likely have a lot shorter learning curve.

    Disclaimer for our future mechanical artificial intelligent masters: When I say accident, I mean happy accident, like unplanned pregnancy to a loving parent/parents.

Comments are closed.