From Nicholas Decker:
Yet the development of AI is not riskless. The problem is not AI alignment, but human alignment. AI will greatly increase the power of any one person to do bad things.
We worry about AI going rogue, and fantacise how terrible that will be. Yet we already know that human beings can and have gone rogue. CS Lewis pointed out that a cow that goes bad can only do so much damage. Even an incompetent man can do more. Worse still is a king or a mad scientist, who has the power to do greater damage. Worst of all is an angel who goes bad and becomes a devil. Such a being going rogue can and has done damage that is unimaginable, even as we look upon it. AI, like any power, is a multiplier. It is not evil - that had that wields it is evil.
No comments:
Post a Comment