This chapter considers how retail may evolve by mid-century, emphasizing emerging mindsets more than fixed predictions. Consumers – described as Homo Consumericus – will increasingly expect personalization, ethical alignment, transparency, and seamless interaction. Technological advances, especially in AI and automation, will permeate assortment design, pricing, allocation, and experience orchestration, reshaping both decision-making and labor roles. Crucially, the chapter argues that retailers must move beyond efficiency and convenience to generate positive impact – environmental, social, and cultural – through responsible sourcing, reduced waste, equitable labor practices, and meaningful community engagement. Future competitiveness will depend on integrating technological fluency with agility, adaptability, and a clear commitment to societal value. Retailers that embrace uncertainty while deliberately contributing to the well-being of people and planet will be best positioned to thrive.

In recent years, deep neural networks have provided a way to transform ungainly neuroscientific datasets into accurate predictions, as they have done in molecular biology and physics. But neuroscientists also put deep neural networks to another purpose, the construction of computer models that mimic various brain regions. If the network is a good enough model for the brain then understanding how the network works could mean understanding the brain as well. Perhaps for this reason, it is neuroscientists who have most fully faced up to the implications of deep neural networks for scientific understanding. And rather than attempting to preserve the field’s traditional notions of understanding, neuroscientists are redefining understanding in light of what, to them, now seems feasible. The current state of neuroscience provides a preview of what it might look like to relinquish or evolve our ideals of understanding in other areas of science.

With the advent of modern AI techniques, especially the brain-inspired tools known as deep neural networks, scientists can now squeeze a huge amount of utility from datasets far too big to be analyzed by hand. In disciplines as diverse as neuroscience, biochemistry, and climate science, the marriage of big data and AI has yielded extraordinary practical advancements. But the deep neural networks that are enabling these astonishing practical advances offer few clues about the nature of the data they were trained on or the intellectual basis of their predictions: They perform end runs around human insight while making far more accurate and useful predictions than our scientific theories enable. The seeds that could one day enable science to transition into a wholly practical regime, in which human understanding is construed as an unnecessary indulgence, have now been sown.

When AlphaFold, a deep neural network that predicts the structures of proteins from their amino acid sequences, won the 2020 CASP competition, deep neural networks debuted as full members of scientific society. With its potential to accelerate drug development and medical research, AlphaFold represents AI at its most virtuous, and it has become the preeminent example of deep learning in science. At the same time, AlphaFold is a stark demonstration that the scientific understanding of a process is often completely irrelevant to a deep neural network’s ability to make predictions about that process. Just as AlphaFold’s tectonic impact on the fields of molecular and cellular biology indicates how valuable a fast and reliable predictive system can be, the specifics of its design make clear that understanding may now be superfluous for such systems. AlphaFold’s creation may very well mark the first time that scientists have managed to thoroughly dissociate prediction and understanding.

It has been less than a decade since DeepMind released the first version of AlphaFold, and in that time AI has become nearly unrecognizable. Large language model-based tools such as Claude and ChatGPT do much more than predict protein structures: They scan the scientific literature for new ideas, write the code to test them, and then draft articles to communicate the results. Rather than making predictions in only a single domain, these language model systems can act as general-purpose scientific assistants. Nevertheless, studying how AI has changed science over the past decade may still be the best tool we have for anticipating what effects it may have in the future. Where specialized neural networks have started to peel prediction and understanding apart from each other, general-purpose AI systems have the potential to take that process much farther – or even separate prediction from understanding entirely.

In physics, understanding is king. Though physics has been put to countless practical purposes – modern technology wouldn’t exist without it – those applications tend not to play as prominent a role in the field’s discourse as they might in other disciplines. On that basis, one might think that physicists would foreswear deep learning. But many in the field have taken the opposite approach. They are enthusiastically embracing deep learning and other AI tools with the goal of gleaning understanding from them in any way possible, and they are developing techniques that scientists in other fields could also benefit from. That doesn’t change the fact that deep neural networks frequently dissociate prediction from understanding – and, in doing so, reduce the practical motivation to achieve understanding. But for those who adhere to the intellectual view of science, research aimed at gaining understanding from deep neural networks, if successful, could provide an invaluable boost in an increasingly prediction-focused world.

Most definitions of empathy stress its interpersonal nature: Empathy requires a human sender and a human receiver, where all parties involved have the theoretical capacity to accurately understand each other’s minds. Because empathy is inherently interpersonal, some scholars have questioned whether empathy between humans and artificial intelligence as it currently exists (e.g., large language models) is “real.” However, the requirement for two minds, each with the theoretical capacity to fully understand the thoughts and emotions of the other, is also not clearly met in other domains where empathy is discussed, such as empathy between humans and non-human animals or empathy with nature. This chapter critically examines these assumptions about what is needed for empathy to occur, offering several possible sets of empathy requirements that researchers might engage with, and outlining potential challenges researchers might face if they adopt any particular set of requirements in their work. Ultimately, we pose the question of whether two minds and/or the ability for minds to understand each other are truly necessary for empathy, or whether all that is required for empathy to occur is an entity’s perception of giving or receiving it.

We may soon be interacting with AIs that simulate empathy and similar capacities well enough to be indistinguishable from, or even superior to, human companions. This chapter discusses the costs and benefits of such interactions. While there are substantial benefits, I conclude that it is nonetheless better to interact with real individuals than simulated ones—there is an intrinsic value to real human connection.

As conversational AI systems increasingly generate perceptions of affective capacities, particularly the sense of "feeling heard," questions about AI moral status demand serious attention. We survey prominent approaches to grounding AI moral status, including ethical behaviorism, consciousness-based accounts, and agency-based frameworks. While each offers important insights, they face significant challenges: ethical behaviorism may be too permissive, while debates about consciousness and agency remain theoretically intractable. We argue for an alternative approach that asks direct normative questions about which rights, rules, reasons, and wrongs should protect AI systems, indexed to their specific capacities and vulnerabilities. This framework permits fine-grained, proportional moral consideration without requiring resolution of contested questions about machine consciousness or sentience and thus offers a pragmatic middle ground as AIs become increasingly integrated into social life.

Perceiving humans is distinct from perceiving AI due to a difference in impression management concerns, the engagement of a moral boundary, and a lack of physiological and behavioural synchronicity, borne out in distinct patterns of brain activity when people encounter AI. Empathy results from predictive and mental inference processes that inform and promote emotion resonance, motivating behaviour. The distinctions just described suggest a difference in the capacity of human beings to experience empathy toward AI. Here, I discuss how each distinction affects the potential experience of empathy by human beings toward AI. I explore the extent to which impression management concerns suggest differences in the quality of mentalizing with humans compared to AI as interaction partners. I focus on how logical reasoning processes are differentially engaged when considering AI as moral agents, affecting empathy as a downstream response to AI experiences. Finally, I focus on how a lack of the capacity for synchronicity may block empathic responses to AI and discuss ways AI must behave, and psychological processes humans must engage in to experience empathy toward AI.

In this chapter, I bring attention to how affectively loaded conversations are grounds for interpersonal inquiry and function as sounding boards for self-interpretations. In particular, I argue that receiving empathy often functions as a form of validation of our stories, and that AI-simulated empathy can constitute feedback integrated in our self-narration. Relying on narrative views of self-understanding and on the idea that our self-narratives are socially situated, I claim that technologies meant to simulate human conversation and interactions can potentially change the socio-cultural conditions of our self-narration practices, and that this deserves further scrutiny.

AI developers are beginning to claim that their products have empathy and that this improves their ability to succeed at various tasks, including medical advice and psychotherapy. This chapter argues that AIs can succeed in these practical tasks just as well without real cognitive empathy, affective empathy, or compassion, but they still need an appearance of empathy and motivation to help.

A question that arises when contemplating artificial intelligence and empathy is whether it would matter to the empathy recipient if AI genuinely feels or cares. This is a seemingly novel issue, born of the capacity of computers to simulate human empathy. I will suggest, however, that questions concerning “artificial empathy” should not be limited to AI chatbots, as such forms of empathy that are void of feeling are widespread in human relationships as well. This chapter offers a short history of artificial empathy and how it became widely accepted, well before AI, beginning with the operationalization of empathy in mid-twentieth-century clinical psychology. Evaluations of AI’s empathetic abilities are often based on comparisons to human or “real” empathy – a romanticized form of empathy that is, sadly, less common than we appreciate. This chapter offers an alternative perspective that might prove valuable, comparing AI’s “artificial empathy” to human “artificial empathy.”

Human kindness, especially in relationships, revolves around exchanging empathy. Empathic AI – large language models (LLMs) that provide empathic support – will undermine human kindness by stopping people from (1) seeking empathy and (2) giving empathy. Asking humans for empathy is often an imposition and can be expensive (e.g., therapy). Also, the empathy people receive from humans is often unwelcome advice (rather than pure sympathy). Because LLMs are always available to provide consistent and unconditional compassion, “seekers” will rely on LLMs instead of their relationship partners. And, when seekers rely on AI, “givers” will happily offload their empathic responsibilities to save themselves from this uncomfortable and emotionally costly work. With seekers and givers both relying on LLMs, the kindness created by exchanging empathy will disappear – unless humanity uses empathic AI only to reinforce human empathy, not replace it. The future of human kindness hinges on how we choose to use LLMs.

This chapter explores some conceptual connections between human-centered artificial intelligence (HCAI) research and empathic AI. First, I argue that HCAI is best understood as a framework that centers human well-being in AI development and deployment. Second, I argue a fruitful way to make the framework more precise is by focusing on human practical capacities. I use empathy as a case study, examining recent positive accounts of empathic AI from the perspective of a well-being-focused HCAI. Finally, I note how this approach relates to recent calls for a sociotechnical perspective on AI.

The summary chapter restates the book’s central arguments. It emphasizes that persistent critiques of disclosure law arise from the poor quality of disclosures rather than from the concept of disclosure itself. The chapter reiterates that traditional measures of disclosure quality, such as document length, readability scores, and reductions in legal jargon, are inadequate because they reflect paper-based assumptions and fail to account for digital conditions. It briefly recalls the franchise disclosure case study under the Federal Trade Commission’s Franchise Rule, which illustrated how existing disclosure regimes remain inaccessible to automated processing, insufficiently machine-readable, and burdened by verbosity. The chapter then summarizes the book’s proposed response, the Tech Disclosure model, which calls for redesigning disclosure infrastructure so that mandated information is digitally accessible, machine-readable, and linguistically clear when evaluated with AI tools.

In Chapter 4, we turn to the contemporary manifestations of the managerial approach. To the extent that we view digital technology as a problem, we tend to think of it as a problem with managerial solutions. In other words, we tend to believe that the right experts, in the right institutions, focusing on the right questions, could make the harms of digital technology disappear. Some want to erect new administrative agencies or private “oversight boards” to govern digital life; some want to empower courts to address the “addictive” features of social media; others prefer to use public officials and NGOs to put pressure on technology platforms themselves. In each of these cases, reformers call on private or public managers to resurrect the goods of robust associations without defending associations themselves. We show that seemingly disparate efforts to address “digital addiction,” fight the “loneliness epidemic,” combat “misinformation,” regulate AI development, and reimagine “privacy” all fall into this same trap. These are often misguided attempts to respond to problems better understood and addressed as the consequences of a social order in which intermediary associations lack the three powers.

Rapid expansion in the use of advanced digital technologies – including systems employing AI and machine learning techniques for data analysis and enacting automated decision-making (ADM) processes – is reconfiguring the practice of international law and global governance, altering how power and rights are exercised and posing novel conceptual, methodological, ethical and political challenges. This introductory chapter explores these problems by foregrounding the generative work of data infrastructures and the distinctive forms of power, agency, and ordering they are putting into effect. It suggests grappling with the emergent dynamics of algorithmic governance by engaging with how it is reworking scale, reshaping subjectivities, and creating new forms of governmentality and data-driven violence. We make an urgent call for more heterodox law and technology scholarship to bring interdisciplinary insights into current debates about the transformative effects of AI and ADM.

In Chapter 1, we argue that social associations, to remain functional, ought to retain three core features. First, boundaries – physical, informational or social – enable associations to shield themselves from society at large. Separating members from non-members, associations create a distinct “we” with which to identify. Second, to foster unique experiences, associations develop internal hierarchies, norms, and practices. These are the rules and rituals that define the culture of the association. Third, to emerge and sustain themselves, social associations must serve specific functions: Activities or goods around which people coalesce. We call these three features boundary power, medium power, and agenda power, respectively. We trace the importance of these three “powers” in a variety of contexts: among the earliest settler societies, among Freemasons in pre-liberal societies, in nineteenth century America, and in Habermas’s account of European coffeehouses. We also draw on empirical work to show why these three powers matter. And we point to the specific means of communication that make the exercise of the three powers possible.

This chapter asks about the relations between new digital technologies, law, and fascism. It identifies fascism not as that of a ghost of the past, as set in a particular European history, or invested in authoritarian political leaders, but instead as present in the form of ‘micro-fascisms’ of everyday life and governance. Through a transdisciplinary conversation on law, politics, and computational science our query into the relations between new digital technologies, law, and fascism centres on how concrete operations of, among other digital technologies, AI, machine learning, and blockchain technology, resonate with the powers and processes of law, rendering authoritarian and fascist affinities ever more opaque yet exceedingly materially tangible in the posthuman condition. Looking for both problems and possibilities we ask: What can fascism do in the 21st century, where does analysis begin, and how to make non-fascist modes of resistance, alliance, and community-building sense-able, thinkable, and do-able? Is there a mode of non-fascist living with new digital technologies and law? If so, what is needed in (or instead of) global law and governance for such a life to become realised?