How can we regulate private power in a globalized, digitized world where state-centered sovereignty, territorial boundaries, and traditional legal frameworks fall short? This introductory chapter provides an overview of the book, its arguments, methodology, and contributions, addressing the urgent need for accountability mechanisms to tame the increasingly unilateral global governance by a handful of corporations. Focusing on content moderation, it examines two key case studies: the EU’s Digital Services Act (DSA) and Meta’s Oversight Board. Both exemplify “emulation,” where public law mechanisms, particularly constitutional and administrative, are adapted to private governance.

Analyzing these “Emulated Guardians”–institutions borrowing the legitimacy of courts while operating in private or hybrid contexts–this book highlights their reliance on performativity and public perception to assert authority. Through interdisciplinary analysis, empirical findings, and expert interviews, the book reveals the ambivalent outcomes of emulation: promising tools for accountability yet sometimes lacking practical efficacy. Ultimately, this work frames these mechanisms as harbingers of new accountability norms, arguing that governance in the digital age demands not only novel institutions but also robust public engagement. It situates these developments within broader debates about power, legitimacy, and the evolving role of public law ideals in globalized, networked environments.

The AI we use is powerful, and its power is increasing rapidly. If this powerful AI is to serve the needs of consumers, voters, and decision-makers, then it is imperative that the AI is accountable. In general, an agent is accountable to a forum if the forum can request information from the agent about its actions, if the forum and the agent can discuss this information, and if the forum can sanction the agent. Unfortunately, in too many cases today’s AI is not accountable – we cannot question it, enter into a discussion with it, let alone sanction it. In this chapter we relate the general definition of accountability to AI, we illustrate what it means for AI to be accountable and unaccountable, and we explore approaches that can improve our chances of living in a world where all AI is accountable to those who are affected by it.

Responsible Artificial Intelligence (RAI) principles are informed by a commitment to attending to local context when designing and deploying Artificial Intelligence (AI) systems. Yet, Artificial Intelligence (AI) systems are developed and deployed across distributed and evolving sociotechnical networks, across many jurisdictions and contexts. Responsible AI practitioners thus face a significant challenge: how to ensure that an AI system is contextually appropriate? That is, how to ensure that AI systems are designed to adapt responsibly to the local social, cultural, legal, and operational settings in which they are used? In this chapter we take up this question and consider practices and opportunities for contextualizing AI during AI development and deployment. In doing so, we draw on Joan Fujimura’s concept of articulation work and illustrate methods for contextualizing AI through a case study of AI evaluation in the cultural heritage sector.

This chapter explores the multifaceted nature of AI transparency. First, a general introduction first presented by Haresamudram et al. is given on what transparency entails in AI, distinguishing between algorithmic, interaction, and social transparency. Algorithmic transparency concerns the technical dimension of AI, including explainability techniques and the disclosure of other technical aspects. Interaction transparency allows users to recognize their interactions with AI systems, while social transparency concerns organizational governance structures. In a second step, we outline challenges to transparency, such as technical complexity, proprietary constraints, and potential conflicts with privacy regulations. Subsequently, this chapter provides an overview of the practical application of transparency, followed by real-world examples of transparency initiatives, highlighting both successes and failures. The final section argues that effective transparency requires strong regulatory frameworks, international standards, and AI literacy among stakeholders to ensure meaningful transparency beyond superficial disclosure.

The primary threat of AI is not a dystopian uprising, but the quiet, rational surrender of human autonomy for convenience. While existing risks like discrimination and disinformation are concerning, another profound danger is one we willingly embrace: a life optimized by AI. As we delegate increasingly consequential decisions – from career paths to personal relationships – to demonstrably superior AI systems, we become meat puppets, happier and healthier, yet stripped of meaningful agency. The chapter argues that this philosophical loss, driven by the competitive pressure to optimize our lives, likely cannot be curbed by regulation and will be the defining, irreversible tragedy of the AI age.

This chapter undertakes a detailed comparison between traditionalism and other theories of constitutional law. It first distinguishes the activity of constitutional interpretation from that of determining what the law of the Constitution is. Next, it compares originalism and traditionalism, and after that nonoriginalism and traditionalism. The upshot is that traditionalism can certainly be consistent with at least some varieties of originalism, but the ends of the theories are different. The same is true of most nonoriginalist theories, which are largely theories of interpretation. The chapter concludes by considering various cases of possible conflict between text and tradition to see which has, historically, tended to trump the other. The historical examples suggest that tradition holds its own in such cases. But the conflict is unlikely to come up very often, for reasons the chapter describes.

This chapter challenges the assumption that placing a “human in the loop” is an adequate safeguard against the harms of AI. It introduces the concept of automation bias – a cognitive tendency for human decision-makers to over-rely on automated systems, often ignoring contradictory evidence or their own training. The chapter demonstrates how the human operator can become a mere “rubber stamp” for a flawed machine.

The risk is magnified with generative AI (GenAI), whose polished, authoritative-sounding outputs and ubiquitous availability deepen trust and erode critical thinking, leading to both errors of commission (following bad advice) and errors of omission (failing to monitor for errors). The chapter concludes that effective mitigation requires more than just human oversight; it demands intentional changes to system design, organizational culture, and human training to foster a healthy skepticism toward AI recommendations.

Chapter 5 traces the regulatory evolution of shareholder – and more broadly, investor – stewardship in the UK, from early investor-led initiatives to the UK Stewardship Code 2020. It begins with the Institutional Shareholders’ Committee, whose statements between 1991 and 2009 gradually reframed activism as stewardship. It then examines the shift to regulator-led stewardship, marked by the Financial Reporting Council’s first-generation UK Stewardship Code (2010/2012), which aimed to foster a market for engagement but remained rooted in shareholder oversight of listed equities. The chapter next assesses the second-generation code (2020), which moves beyond shareholder engagement towards a broader model of investor stewardship. This redefenition embeds ESG considerations, systemic risk, and sustainable value within a principles-based, narrative-driven regime. The analysis also considers the institutional mechanisms supporting implementation and the code’s influence as a global benchmark. While the UK regime has pushed the regulatory frontier, key tensions persist – particularly around enforcement, the translation of normative goals into practice, and the limits of soft law in governing investor conduct.

Chapter 1 reframes how institutional investors exercise power and are held to account in a world shaped by financial intermediation, systemic risk, long-term value concerns, and evolving societal expectations. Tracing the evolution and limits of shareholder governance – including the rise of shareholder activism and the contested promise of shareholder democracy – it introduces the book’s central puzzle: how to institutionalise investor stewardship in ways that are normatively coherent, empirically grounded, and responsive to systemic interdependence. The chapter sets out a model of enlightened shareholder – and more broadly, investor – stewardship, defined by multi-level responsibility, plural accountability, and attention to ‘unseen others’, which reimagines institutional investors as custodians of capital across time, stakeholders, and systems. It outlines the book’s tripartite contribution – conceptual, empirical, and regulatory – and presents its analytical trajectory, interpretive framework, and institutional vision. Anchored in the UK but with global relevance, the chapter sets the stage for rethinking capital’s role in serving public and private interests.