#AoIR2024 Industry

Session Chair: Helen Kennedy
 
Presentation 1
 
Rendering Regulability in AI Supply Chains: Technical and Political Challenges
Robert Gorwa⁽¹⁾, Michael Veale⁽²⁾
1: Berlin Social Science Center (WZB), Germany; 2: University College London
 
The goal of this paper-in-progress is to build upon the emerging interdisciplinary literature on AI supply chains and combine it with an analytical framework and insights gleaned from the extant governance literature on supply chain regulation more generally. We seek to: (a) better understand the novel technical challenges that face AI supply chains from a regulatory perspective, establishing similarities and differences between these new and complex socio-technical 'stacks' in contrast to more widely studied traditional products; (b) survey the political and legal literatures on supply chain governance to pull out some key arguments and hypotheses that may be applicable for the present and future of AI regulation efforts; and (c) offer few brief case studies that explore how emblematic regulatory efforts --- such as certification schemes (Matus and Veale, 2022), standardization efforts (Mueller et al., 2009), norm-building initatives (Boersma, 2018) disclosure requirements (Turner, 2016; Weihrauch et al., 2023), and auditing frameworks (LeBaron et al., 2017) --- have under certain economic and institutional conditions alternatively succeeded and/or fallen short of the mark. In doing so, we emphasise how the arrangement of algorithmic supply chains by leading actors within them can be constructed specifically to make these efforts fail or flop, nonsensically, by making intervention points seem infeasible, trade-offs seem toxic, or relations practically inscrutable.
 
 
Presentation 2
 
Generative AI and the Information Commons: Controversy, Copyright, and Closure
Fenwick McKelvey, Bart Simon, Luciano Frizzera
Concordia University, Canada
 
In 2023, The United States government concluded a public consultation “to examine the copyright law and policy issues raised by artificial intelligence (AI) technology, including the scope of copyright in works generated using AI tools and the use of copyrighted materials in AI training.” The consultation received 10,371 submissions from the public and most large language firms. Drawing on large-scale controversy analysis, we analyze these submissions through topic modelling and named-entity recognition to map key positions as they relate directly or indirectly to matters of the regulation of the commons. Preliminary analysis finds that regulatory submissions that foreclose, or freeze out, the radical challenge of genAI to the commons by treating the governance issue as a copyright question. Large AI firms either argue their efforts amount to fair use, constituting a free-for-all approach to the commons or for cartel like models where elite firms coordinate their privileged access to common-pool resources. Conversely, authors and rights-holders advocated for a transactional commons – akin to a global blockchain – built on contracts and licensing. Our analysis of a key global corpus of genAI governance debates becomes the empirical basis for our emphasis on the commons. Collectively, we argue these efforts collaborate to freeze out the controversiality of genAI in relation to the commons, letting copyright foreclose the more radical challenge of genAI to the imagination and future of the commons.
 
 
Presentation 3
 
Mapping AI Policymaking (2016-2024) in China: Policies, Actors, and Instruments
Xiufeng Jia⁽¹⁾, Lianrui Jia⁽²⁾
1: University of Sussex, United Kingdom; 2: University of Sheffield, United Kingdom
 
Scholars and policymakers have emphasised the urgent need for responsible governance, development, and use of AI. In China, AI adoption permeates sectors from manufacturing and service to medical care and entertainment. The Chinese national government is leveraging AI to drive economic and social advancement, and has opened up its ambition to become a global AI superpower by 2030. China is also making its way to become one of the key governments (e.g., the EU, the US, and the UK) that contributes to the development of international AI standards (Schmitt, 2022). Since China stipulated a national AI action implementation plan in 2016, several recent studies (Cheng and Zeng, 2023; Li and Yang, 2021; Roberts et al., 2023; Roberts et al., 2021; Zeng, 2020) have started reflecting on the status of governance within Chinese contexts. The main framework for examination focuses on AI-related ethical frameworks and authoritarian governance in China. However, given the dynamic and expansive nature of AI policymaking, existing studies tend to adopt a fairly narrow definition of what constitutes AI policy. This overlooks the diversity of policy instruments and their interplay with industrial policies, industry (self-)regulations, and the role of AI users. Therefore, our project aims to systematically map the evolution of AI policymaking in China as the necessary first step to understand the emerging actors, policy patterns, and development paths in China’s approach to AI governance.
 
 
Presentation 4
 
RETHINKING AI FOR GOOD: CRITIQUE, REFRAMING AND ALTERNATIVES
Faranak Hardcastle⁽¹⁾, Sujatha Raman⁽¹⁾, Christer de Silva⁽¹⁾, Jenny Davis⁽²⁾, Ehsan Tavakoli-Nabavi⁽¹⁾
1: Australian National University, Australia.; 2: Vanderbilt University, USA.
 
AI for Social Good (AI4SG) initiatives have emerged in various sectors. However, AI's non-neutral nature challenges claims that the “good” can simply be inferred by association with broad goals such as the Sustainable Development Goals (SDGs). The lack of a clear definition of "the good” or what it entails in practice risks making AI4SG an empty signifier. This ambiguity allows unchecked interventions, undermining societal efforts to align future AI developments with public good. In this article, we adopt a socially situated public good framework from the social studies of quantum technologies proposed by Roberson et al (2021) and use insights from critical AI scholarship to tailor this framework to AI4SG initiatives. Analysing AI4SG initiatives, and building upon existing critical literature, we scrutinize these initiatives with regards to the framings of the research problems, the wider social and institutional context in which AI initiatives are imagined to be applied and used, as well as the wider network of scientists, stakeholders and publics involved in their co-production. We argue that much of the AI4SG literature abstracts AI from social and contextual realities, making it difficult to clarify the ways in which they might in fact have an impact in the world. Outlining our first iteration, we argue that co-creating this framework demands iterative refinement and ongoing dialogue with diverse stakeholders, especially in the Global South.