It is well-known that, to be properly valued, high-quality products must be distinguishable from poor-quality ones. When they are not, indistinguishability creates an asymmetry in information that, in turn, leads to a lemons problem, defined as the market erosion of high-quality products. Although the valuation of generative artificial intelligence (GenAI) systems’ outputs is still largely unknown, preliminary studies show that, all other things being equal, human-made works are evaluated at significantly higher values than machine-enabled ones. Given that these works are often indistinguishable, all the conditions for a lemons problem are present. Against that background, this Chapter proposes a Darwinian reading to highlight how GenAI could potentially lead to “unnatural selection” in the art market—specifically, a competition between human-made and machine-enabled artworks that is not based on the merits but distorted by asymmetrical information. This Chapter proposes solutions ranging from top-down rules of origin to bottom-up signalling. It is argued that both approaches can be employed in copyright law to identify where the human author has exercised the free and creative choices required to meet the criterion of originality, and thus copyrightability.

This chapter will focus on how Chinese and Japanese copyright law balance content owner’s desire for copyright protection with the national policy goal of enabling and promoting technological advancement, in particular in the area of AI-related progress. In discussing this emerging area of law, we will focus mainly on the two most fundamental questions that the widespread adoption of generative AI pose to copyright regulators: (1) does the use and refinement of training data violate copyright law, and (2) who owns a copyright in content produced by or with the help of AI?

This chapter explores the intricate relationship between consumer protection and GenAI. Prominent tools like Bing Chat, ChatGPT4.0, Google’s Gemini (formerly known as Bard), OpenAI’s DALL·E, and Snapchat’s AI chatbot are widely recognized, and they dominate the generative AI landscape. However, numerous smaller, unbranded GenAI tools are embedded within major platforms, often going unrecognized by consumers as AI-driven technology. In particular, the focus of this chapter is the phenomenon of algorithmic consumers, whose interactions with digital tools, including GenAI, have become increasingly dynamic, engaging, and personalized. Indeed, the rise of algorithmic consumers marks a pivotal shift in consumer behaviour, which is now characterized by heightened levels of interactivity and customization.

This chapter introduces a selection of methods applicable for identifying and extracting paradata from existing datasets and data documentation which can then be used to complement existing formal documentation of practices and processes. Data reuse, in its multiple forms, enables researchers to build upon the foundations laid by previous studies. Retrospective methods for eliciting paradata, including qualitative and quantitative backtracking and data forensics, provide means to get insights into past research practices and processes for data-driven analysis. The methods discussed in this chapter enhance understanding of data-related practices and processes, reproducibility of findings by facilitating the replication and verification of results through data reuse. Key references and further reading are provided after each method description.

Generative AI has catapulted into the legal debate through the popular applications ChatGPT, Bard, Dall-E, and others. While the predominant focus has hitherto centred on issues of copyright infringement and regulatory strategies, particularly within the ambit of the AI Act, it is imperative to acknowledge that generative AI also engenders substantial tension with data protection laws. The example of generative AI puts a finger on the sore spot of the contentious relationship between data protection law and machine learning built on the unresolved conflict between the protection of individuals, rooted in fundamental data protection rights and the massive amounts of data required for machine learning, which renders data processing nearly universal. In the case of LLMs, which scrape nearly the whole internet, this training inevitably relies on and possibly even creates personal data under the GDPR. This tension manifests across multiple dimensions, encompassing data subjects’ rights, the foundational principles of data protection, and the fundamental categories of data protection. Drawing on ongoing investigations by data protection authorities in Europe, this paper undertakes a comprehensive analysis of the intricate interplay between generative AI and data protection within the European legal framework.

Research on paradata practices provides diverse insights for the management of paradata. This chapter draws on the existing body of research to inform paradata practices in repository settings including research data archives, repositories and research information management contexts. Four categories of paradata needs (methods; scope; provenance; knowledge representation) are described as well as two major categories of paradata relevant from a repository perspective (core paradata i.e. information commonly perceived as being paradata, and potential paradata i.e. information with potential to function as paradata). Further, the chapter discusses three broad management approaches and a set of intermediary strategies of standardisation and embracing the messiness paradata, and of cultivating paradata literacy to manage different varieties of core paradata and potential paradata.

Making sense of data, and making it useful and manageable, requires understanding of both what the data is about but also where it comes from and how it has been processed, and used. An emerging interdisciplinary corpus of literature terms information about the practices and processes of data making, management and use as paradata. This introductory chapter to a first comprehensive overview of the concept and phenomenon of paradata from data management and knowledge organisation perspectives contextualises the notion and provides an overview of the volume and its aims and starting points.

This chapter provides an outline analysis of the evolving governance framework for Artificial Intelligence (AI) in the island city-state of Singapore. In broad terms, Singapore’s signature approach to AI Governance reflects its governance culture more broadly, which harnesses the productive energy of free-market capitalism contained within clear guardrails, as well as the dual nature (as a regulator and development authority) of Singapore’s lead public agency in AI policy formulation. Singapore’s approach is interesting for other jurisdictions in the region and around the world and it can already be observed to have influenced the recent Association of South East Asian Nations (ASEAN) Guide on AI Governance and Ethics which was promulgated in early 2024.

This chapter explores the privacy challenges posed by generative AI and argues for a fundamental rethinking of privacy governance frameworks in response. It examines the technical characteristics and capabilities of generative AIs that amplify existing privacy risks and introduce new challenges, including nonconsensual data extraction, data leakage and re-identification, inferential profiling, synthetic media generation, and algorithmic bias. It surveys the current landscape of U.S. privacy law and its shortcomings in addressing these emergent issues, highlighting the limitations of a patchwork approach to privacy regulation, the overreliance on notice and choice, the barriers to transparency and accountability, and the inadequacy of individual rights and recourse. The chapter outlines critical elements of a new paradigm for generative AI privacy governance that recognizes collective and systemic privacy harms, institutes proactive measures, and imposes precautionary safeguards, emphasizing the need to recognize privacy as a public good and collective responsibility. The analysis concludes by discussing the political, legal, and cultural obstacles to regulatory reform in the United States, most notably the polarization that prevents the enactment of comprehensive federal privacy legislation, the strong commitment to free speech under the First Amendment, and the “permissionless” innovation approach that has historically characterized U.S. technology policy.

This chapter introduces methods for generating and documenting paradata before and during data creation practices and processes (i.e. prospective and in-situ approaches, respectively). It introduces formal metadata-based paradata documentation using standards and controlled vocabularies to contribute to paradata consistency and interoperability. Narrative descriptions and recordings are advantageous for providing contextual richness and detailed documentation of data generation processes. Logging methods, including log files and blockchain technology, allow for automatic paradata generation and for maintaining the integrity of the record. Data management plans and registered reports are examples of measures to prospectively generate potential paradata on forthcoming activities. Finally, facilitative workflow-based approaches are introduced for step-by-step modelling of practices and processes. Rather than suggesting that a single approach to generating and documenting paradata will suffice, we encourage users to consider a selective combination of approaches, facilitated by adequate institutional resources, technical and subject expertise, to enhance the understanding, transparency, reproducibility and credibility of paradata describing practices and processes.

The chapter examines the legal regulation and governance of ‘generative AI,’ ‘foundation AI,’ ‘large language models’ (LLMs), and the ‘general-purpose’ AI models of the AI Act. Attention is drawn to two potential sorcerer’s apprentices, namely, in the spirit of J. W. Goethe’s poem, people who were unable to control a situation they created. Focus is on developers and producers of such technologies, such as LLMs that bring about risks of discrimination and information hazards, malicious uses and environmental harms; furthermore, the analysis dwells on the normative attempt of EU legislators to govern misuses and overuses of LLMs with the AI Act. Scholars, private companies, and organisations have stressed limits of such normative attempts. In addition to issues of competitiveness and legal certainty, bureaucratic burdens and standard development, the threat is the over-frequent revision of the law to tackle advancements of technology. The chapter illustrates this threat since the inception of the AI Act and recommends some ways in which the law has not to be continuously amended to address the challenges of technological innovation.

Generative AI offers a new lever for re-enchanting public administration, with the potential to contribute to a turning point in the project to ‘reinvent government’ through technology. Its deployment and use in public administration raise the question of its regulation. Adopting an empirical perspective, this chapter analyses how the United States of America and the European Union have regulated the deployment and use of this technology within their administrations. This transatlantic perspective is justified by the fact that these two entities have been very quick to regulate the issue of the deployment and use of this technology within their administrations. They are also considered to be emblematic actors in the regulation of AI. Finally, they share a common basis in terms of public law, namely their adherence to the rule of law. In this context, the chapter highlights four regulatory approaches to regulating the development and use of generative AI in public administration: command and control, the risk-based approach, the experimental approach, and the management-based approach. It also highlights the main legal issues raised by the use of such technology in public administration and the key administrative principles and values that need to be safeguarded.

The rapid development of generative artificial intelligence (AI) systems, particularly those fuelled by increasingly advanced large language models (LLMs), has raised concerns of their potential risks among policymakers globally. In July 2023, Chinese regulators enacted the Interim Measures for the Management of Generative AI Services (“the Measures”). The Measures aim to mitigate various risks associated with public-facing generative AI services, particularly those concerning content safety and security. At the same time, Chinese regulators are seeking the further development and application of such technology across diverse industries. Tensions between these policy objectives are reflected in the provisions of the Measures that entail different types of obligations on generative AI service providers. Such tensions present significant challenges for implementation of the regulation. As Beijing moves towards establishing a comprehensive legal framework for AI governance, legislators will need to further clarify and balance the responsibilities of diverse stakeholders.

Generative artificial intelligence (GenAI) raises ethical and social challenges that can be examined according to a normative and an epistemological approach. The normative approach, increasingly adopted by European institutions, identifies the pros and cons of technological advancement. The main pros concern technological innovation, economic development and the achievement of social goals and values. The disadvantages mainly concern cases of abuse, use or underuse of Gen AI. The epistemological approach investigates the specific way in which Gen AI produces information, knowledge, and a representation of reality that differs from that of human beings. To fully realise the impact of Gen AI, our paper contends that both these approaches should be pursued: an identification of the risks and opportunities of Gen AI also depends on considering how this form of AI works from an epistemological viewpoint and our ability to interact with it. Our analysis compares the epistemology of Gen AI with that of law, to highlight four problematic issues in terms of: (i) qualification; (ii) reliability; (iii) pluralism and novelty; (iv) technological dependence. The epistemological analysis of these issues leads to a better framing of the social and ethical aspects resulting from the use, abuse or underuse of Gen AI.

Drawing on the extensive history of study of the terms and conditions (T&Cs) and privacy policies of social media companies, this paper reports the results of pilot empirical work conducted in January-March 2023, in which T&Cs were mapped across a representative sample of generative AI providers as well as some downstream deployers. Our study looked at providers of multiple modes of output (text, image, etc.), small and large sizes, and varying countries of origin. Our early findings indicate the emergence of a “platformisation paradigm”, in which providers of generative AI attempt to position themselves as neutral intermediaries similarly to search and social media platforms, but without the governance increasingly imposed on these actors, and in contradiction to their function as content generators rather than mere hosts for third party content.