This chapter is about a refinement acting engine (RAE) used on a hierarchical task-oriented representation. It relies on an expressive, general-purpose language that offers rich programming control structures for online decision-making. A collection of refinement methods describes alternative ways to handle tasks and react to events. A method can be any complex algorithm, decomposing a task into subtasks and primitive actions. Subtasks are refined recursively. Nondeterministic actions trigger sensory-motor procedures that query and change the world nondeterministically. We assume that the methods are manually specified and that RAE chooses the appropriate method for the task and context at hand heuristically.

The recent developments of large language models (LLMs) and their extension in multimodal foundation models have introduced new perspectives in AI. An LLM is basically a very large neural net trained as a statistical predictor of the likely continuation of a sequence of words. LLMs have excellent competencies over a broad set of NLP tasks. Additionally, LLMs demonstrate the emergence of deliberation capabilities for reasoning, common sense, problem solving, code writing, and planning. These abilities have not been designed for in LLMs. They are unexpected and remain to a large extent poorly understood. Although error-prone and imperfect, they open up promising perspectives for acting, planning, and learning, which are presented in this chapter.

This chapter is about domain-independent classical-planning algorithms, which until recently were the most widely studied class of AI planning algorithms. The chapter classifies and describes a variety of forward search, backward search, and plan-space planning algorithms, as well as heuristics for guiding the algorithms.

This chapter sets the foundation for the next two chapters. It introduces the reader to robotics platforms for the development of acting, planning, and learning functions. The study of motion is based on classical mechanics for the modeling of forces and their effects on mouvements. Robotics builds on this knowledge to master computational motion, navigation, and manipulation over different types of devices and environments. Robotic devices are informally introduced in the following section. Motion problems and the metric representations with continuous state variables needed for geometric, kinematic, and dynamic operational models are then presented. Section 20.3 introduces localization and navigation problems, followed by a section on manipulation problems and their representations.

This chapter is about representing HTN planning domains and solving HTN planning problems. Several of the formal definitions require the same "classical planning" restrictions as in Part I, but most practical HTN implementations loosen or drop several of these restrictions. We first discuss ways to represent and solve planning problems in which there is a totally ordered sequence of tasks to accomplish. We then generalize to allow partially ordered tasks and describe ways to combine classical planning and HTN planning. Finally, we briefly discuss heuristic functions, expressivity, and computational complexity.

The chapters in Part II are about algorithms for planning, acting, and learning using hierarchical task networks (HTNs). HTNs can describe ways to perform complex tasks without the overhead of searching through a large state space, how to avoid situations where unanticipated events are likely to cause bad outcomes, and how to recover when unanticipated events occur.

This chapter discusses several ways for actors to use HTN domain models. These include a way to use HTN methods for purely reactive acting, some simple ways for an actor to make use of an HTN planner, and some ways to repair HTN plans when unexpected events occur during acting.

The hierarchical refinement approach in the previous two chapters requires a priori domain knowledge of the methods, action models, and heuristics used by RAE and UPOM. The topic of this chapter is to use machine learning techniques to synthesize planning heuristics and domain knowledge. It illustrates the "planning to learn" paradigm for learning domain-dependent heuristics to guide RAE and UPOM. Given methods and a sample function, UPOM generates near-optimal choices that are taken as targets by a deep Q-learning procedure. The chapter shows how to synthesize methods for tasks using hierarchical reinforcement techniques.

This chapter introduces social scientific perspectives and methods applicable to observing the relationship between artificial intelligence (AI) and religion. It discusses the contributions that anthropological and sociological approaches can make to this entanglement of two modern social phenomena while also drawing attention to the inherent biases and perspectives that both fields bring with them due to their histories. Examples of research on religion and AI are highlighted, especially when they demonstrate agile and new methodologies for engaging with AI in its many applications; including but not limited to online worlds, multimedia formats, games, social media and the new spaces made by technological innovation such as the innovations such as the platforms underpinning the gig economy. All these AI-enabled spaces can be entangled with religious and spiritual conceptions of the world. This chapter also aims to expand upon the relationship between AI and religion as it is perceived as a general concept or object within human society and civilisation. It explains how both anthropology and sociology can provide frameworks for conceptualising that relationship and give us ways to account for our narratives of secularisation – informed by AI development – that see religion as a remnant of a prior, less rational stage of human civilisation.

This chapter explores the intersection of Hindu philosophy and practice with the development of artificial intelligence (AI). The chapter first introduces aspects of technological growth in Hindu contexts, including the reception of ‘Western’ ideas about AI in Hindu communities before describing key elements of the Hindu traditions. It then shows how AI technologies can be conceived of from a Hindu perspective and moves from there to the philosophical contributions Hinduism offers for global reflection on AI. Specifically, the chapter describes openings and contentions for AI in Hindu rituals. The focus is the use of robotics and/or AI in Hindu pūjā (worship of gods) and the key practice of darśan (mutual seeing) with the divine. Subsequently, the chapter investigates how Hindu philosophers have engaged the distinctive qualities of human beings and their investigation into body, minds and consciousness/awareness. The chapter concludes by raising questions for future research.

Artificial intelligence (AI) is presented as a portal to more liberative realities, but its broad implications for society and certain groups in particular require more critical examination. This chapter takes a specifically Black theological perspective to consider the scepticism within Black communities around narrow applications of AI as well as the more speculative ideas about these technologies, for example general AI. Black theology’s perpetual push towards Black liberation, combined with womanism’s invitation to participate in processes that reconstitute Black quality of life, have perfectly situated Black theological thought for discourse around artificial intelligence. Moreover, there are four particular categories where Black theologians and religious scholars have already broken ground and might be helpful to religious discourse concerning Blackness and AI. Those areas are: white supremacy, surveillance and policing, consciousness and God. This chapter encounters several scholars and perspectives within the field of Black theology and points to potential avenues for future theological areas of concern and exploration.

While we call programs that are new and exciting ‘artificial intelligence’ (AI), the ultimate goal – to produce an artificial general intelligence that can equal to human intelligence – always seems to be in the future. AI can, thus, be viewed as a millenarian project. Groups predicting the second coming of Christ or some other form of salvation have flourished in times of societal stress, as they promise a solution to current problems that is delivered from outside. Today, we project both our hopes and our fears onto AI. Utopian visions range from the personally soteriological prospect of uploading our brains to a vision of a world in which AI has found solutions to our problems. Dystopian scenarios involve the creation of a superintelligent AI that slips from our control or is used as a weapon by malicious actors. Will AI save us or destroy us? Probably neither, but as we shape the trajectory of its future, we also shape our own.