ABSTRACT

After a number of AI-winters, AI is back with a boom. There are concerns that it will disrupt society. The immediate concern is whether labor can win a ‘race against the robots’ and the longer-term concern is whether an artificial general intelligence (super-intelligence) can be controlled. This paper describes the nature and context of these concerns, reviews the current state of the empirical and theoretical literature in economics on the impact of AI on jobs and inequality, and discusses the challenge of AI arms races. It is concluded that despite the media hype neither massive jobs losses nor a ‘Singularity’ is imminent. In part, this is because current AI, based on deep learning, is expensive and difficult for most businesses to adopt, not only displaces but in fact also create jobs, and may not be the route to a super-intelligence. Thus AI is unlikely to have either Utopian or apocalyptic impacts soon. Considering Amara's Law, one should however be wary not to underestimate the long-run impacts of AI.

摘要

在经历了许多AI冬季之后，AI蓬勃发展。有人担心它将破坏社会。当务之急是劳力能否赢得“与机器人的竞争”，而更长期的关注是人工通用情报（超级智能）是否可以得到控制。本文描述了这些担忧的性质和背景，回顾了经济学中关于AI对工作和不平等的影响的经验和理论文献的现状，并讨论了AI军备竞赛的挑战。可以得出结论，尽管媒体大肆宣传，但大量的工作机会流失和“奇异性”都不是迫在眉睫。在某种程度上，这是因为当前基于深度学习的AI昂贵且对大多数企业而言难以采用，不仅取代了，而且实际上还创造了就业机会，并且可能不是通往超级智能的途径。因此，人工智能不太可能很快对乌托邦或世界末日产生影响。考虑中但是，应谨慎对待《阿马拉法则》，不要低估AI的长期影响。