AI Governance

Author:

Dafoe Allan1

Affiliation:

1. Long-term Strategy and Governance, DeepMind

Abstract

Abstract Artificial intelligence (AI) will be a transformative technology, with extreme potential risks and benefits. AI governance refers to the norms and institutions shaping how AI is built and deployed, as well as the policy and research efforts to make it go well. This chapter argues that the field of AI governance should have an expansive and ambitious scope, commensurate to the challenges, with robust internal collaboration given transferable lessons and shared policy opportunities. To make sense of the impacts of AI, the chapter offers three theoretical lenses, focusing on distinct mechanisms, impacts, and challenges. These lenses regard AI as a general purpose technology, an information technology, and an intelligence technology. The chapter then provides a lens on governance focusing on institutional fit and adaptation to the externalities produced by AI. Institutional adaptation will be especially difficult when a governance issue touches on deep social conflicts. Great power security competition poses a particular challenge because it can induce extreme—even existential—risks and is among the hardest dynamics to govern. Building strong competent global institutions to govern powerful AI would be a historically unparalleled challenge, but ultimately may be required to steer away from the greatest risks inherent to great power competition.

Publisher

Oxford University Press

Reference89 articles.

1. Acemoglu, D. (2022). Harms of AI. In J. Bullock (Ed.), The Oxford Handbook of AI Governance. Oxford University Press, 2022.

2. Conceptual models and the Cuban missile crisis.;American Political Science Review

3. Amodei, D., Olah, C., Steinhardt, J., Christiano, P., Schulman, J., & Mané, D. (2016, July). Concrete problems in AI safety. arXiv. https://arxiv.org/abs/1606.06565.

4. Angwin, J., Larson, J., Mattu, S., & Kirchner, L. (2016). Machine bias risk assessments in criminal sentencing. ProPublica, May. https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing.

Cited by 3 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3