Generative AI, generating crisis: Framing opportunity and threat in AI governance
生成式人工智能，生成的危机：人工智能治理中的机遇与威胁框架
Author: Sarah Cheung（张莎拉）
Source：《Information Communication & Society》2025
Abstract: The public release of high-profile generative AI models in late 2022 and early 2023 spurred widespread concerns over 'existential threats' to humanity from AI. This paper explores how this AI 'crisis' impacted ongoing approaches to AI, altering policymakers' strategies to govern these technologies. Drawing on policy analysis approaches that highlight the role of ideas, this paper finds that, in response to perceived crisis and a frame of threat to humanity's public interest, the UK shifted their governance approach to AI. It suggests that digital agenda-setting power in the policymaking arena is emerging as a distinct site of power in the digital era - one with the potential to mitigate or exacerbate the myriad effects from using digital technologies. Exploring AI policymaking as a site of contestation, between actors, ideas and interests, supports understandings of dimensions of power and control in the digital era.
文摘：2022年底至2023年初，备受瞩目的生成式人工智能模型的公开发布，引发了关于人工智能对人类构成“生存威胁”的广泛担忧。本文探讨了这场人工智能“危机”如何影响现行的人工智能应对路径，并改变了政策制定者治理这些技术的策略。借鉴强调观念作用的政策分析方法，本文发现，为了应对感知到的危机以及威胁人类公共利益的框架，英国调整了其人工智能治理路径。研究表明，政策制定领域中的数字议程设置权正作为数字时代一种独特的权力场域显现出来——它具有缓解或加剧使用数字技术所带来的种种影响的潜力。将人工智能政策制定视为行动者、观念和利益之间的博弈场域进行探索，有助于理解数字时代的权力和控制维度。
Keywords: generative AI; AI governance; policy frames; public interest
编译：任艳林、刘鑫
