This chapter aims to present the potential challenges and opportunities brought forth by the integration of care robots within the home- and healthcare services within the Nordic context, stressing the difference between home and conventional or institutional care. In this chapter, we specifically focus on Social and Assistive Robots. Similarly, the chapter aims to discuss if and to what extent these care robots need to be designed for a diversity of users. The chapter suggests Universal Design and accessibility of robots as a potential approach to achieving the care robots’ characteristics of becoming inclusive. Nevertheless, the chapter argues that an inclusive robot approach is important in order to protect the right to health(care), especially if robots will be integrated as part of the home- and/or healthcare services. The chapter exemplifies the need for universal design and accessible robots, as well as the idea of inclusive care robots through empirical work based on a series of interview sessions. Finally, the chapter concludes that: Social and Assistive robots need to be seen in context if such robots are integrated into home- and healthcare services; and that a universal design approach could offer an ethical design charter, a solution, or a guide for designing robots that cater to diverse populations with various situated abilities, considering rights such as the right to healthcare.

Autonomous robots and Artificial Intelligence (AI) are increasingly involved in the commission of criminal offenses, resulting in questions as to who is accountable for such crimes and how human–machine interactions influence criminal responsibility. This chapter first elaborates upon the conditions of machine responsibility and explains why technical systems cannot be granted personhood under today’s criminal law. It then discusses the challenges that complex human–machine interactions bring in terms of the attribution of criminal responsibility, before outlining options regarding how they might be met. Human–machine interactions should be understood as a form of distributed agency. It is therefore essential to define what a legitimate delegation of agency to a machine is. As this chapter shows, this could lead to a modified understanding of the attribution of action under criminal law, or even new norms specifically designed to address automation. Ensuring accountability under criminal law for robots and AI means holding actors accountable who cede agency to such technologies in bad or unjustified ways.

Social norms, often described as the “cement” or “grammar” of society, guide human behavior and infuse it with social meaning. In this chapter, we offer a legal perspective on whether robots should be required to adhere to social norms in the context of human–robot interactions (HRIs). To do so, we examine how the mass introduction of social robots may affect existing norms in different contexts, develop a taxonomy of HRIs as they pertain to social norms, and offer a legal-normative framework of analysis to determine whether and to what extent should robots be legally required to adhere to social norms.

Recent progress has been made towards developing automated companions for the elderly. Building on work in the early days of artificial intelligence that showed that computers could deliver non-directive counselling, the possibility arises that computers could be used to provide people with an opportunity for spiritual conversation. Research using Wizard-of-Oz methodology shows that at least some people find it helpful to have spiritual conversations with what they believe to be an avatar, and work using GPT-3 shows that computers can be an acceptable interlocutor in spiritual conversation. The possibility now arises of developing a spiritual companion that would be personalised for a particular individual and become familiar with their spiritual life. This would not, in every way, replace a human spiritual guide, but could provide a resource that at least some people would find valuable and would assist in their spiritual development.

Chapter 9 offers a comprehensive exploration of decentralized autonomous organizations (DAOs) and their role in the formation of Web3 organizations. It begins by defining DAOs, tracing their history, and outlining the benefits they bring to the digital world. The chapter then delves into the various types of DAOs, including protocol DAOs such as MakerDAO and Lido, charity DAOs such as GitcoinDAO and Ukraine DAO, and investment DAOs, represented by The LAO and MetaCartel Ventures. The chapter also examines Art Collection, Social, Fans, Sports, Media, Tool, and multipurpose DAOs, with examples such as Aragon DAO and Coordinape DAO. However, it is not without noting the disadvantages of DAOs, such as slow decision-making, low active participation rates, legal and tax issues, centralization risk, and vulnerability to flash loan attacks. The future of DAOs is also a topic of discussion, with speculations on how they could be combined with artificial intelligence (AI), attract large institutional investments, or be endorsed by big brands. Furthermore, the chapter considers the possibility of DAOs functioning as virtual countries and their potential role in inclusive finance. The chapter provides a summary, followed by a series of thought-provoking questions for further reflection.

Chapter 9 is devoted to evaluation methods for an important category of classical learning paradigms left out of Chapter 8 so as to receive fuller coverage: unsupervised learning. In this chapter, a number of different unsupervised learning schemes are considered and their evaluation discussed. The particular tasks considered are clustering and hierarchical clustering, dimensionality reduction, latent variable modeling, and generative models including probabilistic PCA, variational autoencoders, and GANs. Evaluation methodology is discussed discussed for each of these tasks.

Chapter 4 examines the scaling of Web3 to support widespread adoption. It first discusses why scalability is crucial for Web3, enabling it to handle high transaction volumes and users such as centralized systems. Scalability refers to the ability to sustain performance amid growth. Key factors are number of users, response time, storage, transaction costs, and throughput. Scalability is challenging due to the blockchain trilemma of decentralization, security, and scalability. Solutions involve optimizations at the network layer (layer 0), blockchain layer (layer 1), and Off-chain layer (layer 2). Layer 0 focuses on data transfer, using protocols such as BloXroute. Layer 1 aims to improve the blockchain itself via methods like sharding or new consensus algorithms. Layer 2 leverages Off-chain processing via rollups, sidechains, and state channels. Each layer has trade-offs. Rollups bundle transactions Off-chain using zero-knowledge proofs or fraud proofs before validating On-chain. Sidechains process transactions externally to relieve the main chain’s load. State channels allow Off-chain transfers between participants. No single scaling approach fits all cases. A combination of solutions across layers tailored to the application offers the most potential.

Chapter 11 completes the discussion of Chapter 10 by raising the question of how to practice machine learning in a responsible manner. It describes the dangers of data bias, and surveys data bias detection and mitigation methods; it lists the benefits of explainability and discusses techniques, such as LIME and SHAP, that have been proposed to explain the decisions made by opaque models; it underlines the risks of discrimination and discusses how to enhance fairness and prevent discrimination in machine learning algorithms. The issues of privacy and security are then presented, and the need to practice human-centered machine learning emphasized. The chapter concludes with the important issues of repeatability, reproducibility, and replicability in machine learning.

Chapter 1 discusses the motivation for the book and the rationale for its organization into four parts: preliminary considerations, evaluation for classification, evaluation in other settings, and evaluation from a practical perspective. In more detail, the first part provides the statistical tools necessary for evaluation and reviews the main machine learning principles as well as frequently used evaluation practices. The second part discusses the most common setting in which machine learning evaluation has been applied: classification. The third part extends the discussion to other paradigms such as multi-label classification, regression analysis, data stream mining, and unsupervised learning. The fourth part broadens the conversation by moving it from the laboratory setting to the practical setting, specifically discussing issues of robustness and responsible deployment.

Simulating religion through computer modelling can demonstrate how fragmentary theories relate, untangle individual lines of causal influence, identify the relative importance of causal factors and enable experimentation that would never be possible (or ethical) in the real world. This chapter reviews the application of computational modelling and simulation to religion, presents findings from specific simulation studies and discusses some of the philosophical issues raised by this type of research. Social simulations are artificial complex systems that we can use to study real-world complex systems. The best of these simulation models are carefully validated in relation to real-world data. Multilevel validation justifies confidence that the causal architecture of the simulation reflects real-world causal processes, thereby delivering an invaluable proxy system into the hands of researchers who study religion.

This chapter explores design guidelines and potential regulatory issues that could be associated with future baby robot interaction. We coin the term “robot natives,” which we define as the first generation of human’s regularly interacting with robots in domestic environments. This term includes babies (0–1 year old) and toddlers (1–3 years old) born in the 2020s. Drawing from the experience of other interactive technologies becoming widely available in the home and the positive and negative impact they have on humans; we propose some insights into the design of future scenarios for baby–robot interaction, aiming to influence future legislation regulating service robots and social robots used with robot natives. Similarly, we aim to inform designers and developers to inhibit robot designs which can negatively affect the long-term interactions for the robot natives. We conclude that a qualitative, multidisciplinary, ethical, human-centered design approach should be beneficial to guide the design and use of robots in the home and around families as this is currently not a common approach in the design of studies in child robot interaction.

Romance between a human and robot will pose many questions for the laws that apply to human–robot interaction and, in particular, family law. Such questions include whether humans and robots can marry and what a subsequent divorce might look like. This chapter considers these issues, organized to track the seasons of romantic relationships, such as cohabitation, engagement, and marriage. Given that marriage is no longer devoid of the possibility of divorce, this chapter also considers issues of property division, alimony, child custody, and child support when a marriage between a human and robot dissolves. Even for skeptics of such a future, given rapid advances in robotics, the applicability of family law to relationships between a human and robot is nonetheless an increasingly relevant thought experiment and intersects with other emerging areas of law, technology, and robotics.

Chapter 8 introduces evaluation procedures for paradigms other than classification. In particular, it discusses evaluation for classical problems such as regression analysis, time-series analysis, outlier detection, and reinforcement learning, along with evaluation approaches for newer tasks such as positive-unlabelled classification, ordinal classification, multi-labeled classification, image segmentation, text generation, data stream mining, and lifelong learning.