Pause artificial intelligence research? Understanding AI policy challenges

被引:2
|
作者
Goldfarb, Avi [1 ]
机构
[1] Univ Toronto, Rotman Sch Management, Toronto, ON, Canada
来源
CANADIAN JOURNAL OF ECONOMICS-REVUE CANADIENNE D ECONOMIQUE | 2024年 / 57卷 / 02期
关键词
HISTORICAL-PERSPECTIVE; PRODUCTIVITY GROWTH; TECHNOLOGY; AUTOMATION; INEQUALITY; SKILLS; TASKS; WAGES; RISE;
D O I
10.1111/caje.12705
中图分类号
F [经济];
学科分类号
02 ;
摘要
Artificial intelligence (AI) may be the next general purpose technology. General purpose technologies, such as the steam engine and computing, can have an outsized impact on productivity through a positive feedback loop between producing and application industries. Along with the discussion of AI's potential to improve productivity come a number of policy concerns related to AI's potential to automate jobs and to create existential risk for humanity. Because of these worries, in March 2023, a widely circulated petition called for a pause in AI research. That letter asked several questions about AI's potential impact on society. This paper examines those questions through an economic lens. It highlights reasons to be optimistic about the long-run impact of AI, while underscoring short-run risks. Economic models provide an understanding of where the ambiguity lies and where it does not. Our models suggest no ambiguity on whether there will be jobs and little ambiguity on long-term productivity growth if AI diffuses widely. In contrast, there is substantial ambiguity on the implications of AI's diffusion for inequality.
引用
收藏
页码:363 / 377
页数:15
相关论文
共 50 条