Artificial Intelligence Needs a Leash, Not a Conscience

The safety of artificial intelligence may be less a problem of ethics — teaching machines how to behave — than one of engineering: AI systems can never be fully trusted, so they must be heavily constrained. Read Full Article »