Gray goo is a more likely outcome on the way to creating "strong AI" that we should be more clearly worried about. A "strong AI" which can perhaps be reasoned with seems significantly less dangerous than accidentally creating a machine designed to optimize something which could conflict with human interests.
Asimov's laws are very interesting but in practice a system would have to have a framework in which to express the concepts.
Yet another reason we need to get DNA replicating somewhere far from this silly rock.
Gray goo is a more likely outcome on the way to creating "strong AI" that we should be more clearly worried about. A "strong AI" which can perhaps be reasoned with seems significantly less dangerous than accidentally creating a machine designed to optimize something which could conflict with human interests. Asimov's laws are very interesting but in practice a system would have to have a framework in which to express the concepts. Yet another reason we need to get DNA replicating somewhere far from this silly rock.