Scholarship

摘要（出版方所载原文摘要） Political scientists have had remarkably little to say about artificial intelligence (AI), perhaps because they are dissuaded by its technical complexity and by current debates about whether AI might emulate, outstrip, or replace individual human intelligence. They ought to consider AI in terms of its relationship with governance. Existing large-scale systems of governance such as markets, bureaucracy, and democracy make complex human relations tractable, albeit with some loss of information. AI’s major political consequences can be considered under two headings. First, we may treat AI as a technology of governance, asking how AI’s capacities to classify information at scale affect markets, bureaucracy, and democracy. Second, we might treat AI as an emerging form of governance in its own right, with its own particular mechanisms of representation and coordination. These two perspectives reveal new questions for political scientists, encouraging them to reconsider the boundaries of their discipline. ...

摘要（出版方所载原文摘要） The rapid and widespread diffusion of generative artificial intelligence (AI) has unlocked new capabilities and changed how content and services are created, shared, and consumed. This special issue builds on the 2021 Policy and Society special issue on the governance of AI by focusing on the legal, organizational, political, regulatory, and social challenges of governing generative AI. This introductory article lays the foundation for understanding generative AI and underscores its key risks, including hallucination, jailbreaking, data training and validation issues, sensitive information leakage, opacity, control challenges, and design and implementation risks. It then examines the governance challenges of generative AI, such as data governance, intellectual property concerns, bias amplification, privacy violations, misinformation, fraud, societal impacts, power imbalances, limited public engagement, public sector challenges, and the need for international cooperation. The article then highlights a comprehensive framework to govern generative AI, emphasizing the need for adaptive, participatory, and proactive approaches. The articles in this special issue stress the urgency of developing innovative and inclusive approaches to ensure that generative AI development is aligned with societal values. They explore the need for adaptation of data governance and intellectual property laws, propose a complexity-based approach for responsible governance, analyze how the dominance of Big Tech is exacerbated by generative AI developments and how this affects policy processes, highlight the shortcomings of technocratic governance and the need for broader stakeholder participation, propose new regulatory frameworks informed by AI safety research and learning from other industries, and highlight the societal impacts of generative AI. ...

摘要（出版方所载原文摘要） It seemed well established that producing a smart device could not, by itself, render someone a personal data controller in the absence of subsequent influence over the processing operations (the influence thesis). In contrast, legal scholars have introduced a new interpretation of European data protection law that seeks to apply the General Data Protection Regulation (GDPR) to the processing operations of smart devices even if no entity influences the processing remotely after the release of the product. This approach classifies producers as personal data controllers for device-based processing (producer-controller thesis). The proponents of the producer-controller thesis highlight the increasing importance of smart devices that store data locally and the need for protecting consumers’ rights in that context. However, as this paper claims, the GDPR is not the proper legal instrument for addressing the safety standards of smart products that process data locally. These considerations relate to legislative texts that prescribe product requirements, such as the AI Act and the Cyber Resilience Act. On those grounds, the present work criticises the producer-controller thesis. As this paper concludes, expanding the concept of ‘controller’ to encompass producers of smart devices does not enhance the protection of the data subjects and does not fit within the current data protection framework of the European Union. ...

摘要（出版方所载原文摘要） Synthetic data – algorithmically generated data – has been considered a novel solution to the data scarcity issue, and a ‘technical fix’ able to fill the gap in areas where real data is sensitive or biased. Different narratives about the nature of synthetic data as either mirroring or replacing real data, alongside diverse evaluation metrics for measuring the fidelity and utility of such data, have proliferated across the machine learning fairness community, in public policy research, privacy and data protection studies, and critical data scholarship. Yet, there is still no consensus on what constitutes ‘high-quality’ synthetic data. Against this background, I demonstrate how the concept of synthetic data introduces an analogical perspective on data. This perspective is relational and regulative, extending the discussion on data quality to encompass questions of data justice and responsible innovation. It invites critical reflections on the purpose and trade-offs involved in synthetic data generation and use, the social practices and power dynamics that underpin and configure it, and how its direction can be shaped in response to changing real-world circumstances and emerging human values. Building on this analysis, I argue that the generation and use of meaningful synthetic data require promoting responsibility in complex AI and data innovation ecosystems, and facilitating forms of algorithmic reparation and responsiveness. ...

摘要（出版方所载原文摘要） This paper is a comparative analysis of how two leading developing nations, China and India, are proposing to regulate artificial intelligence (AI) systems. Despite similarity in circumstances as large developing economies aiming to upgrade their technology sectors and create jobs, the two countries have taken significantly different approaches to AI regulation. Based on interest group theory, we argue that contrasting problem definitions—predominantly in terms of economic competitiveness and national security in China and as applications in India—resulted in the recruitment of very different decision-making groups in the two countries; homogeneous groups of technocrats and security specialists in China and a broader group including consumer advocates in India. This in turn resulted in ambitious and deep policy changes in China and relatively incremental and consensus-based moves in India. ...