Paul Smart

Burden is a key concept in healthcare research, reflecting the challenges that illness and its management impose on patients, caregivers, and healthcare systems. While burden has been the focus of considerable scientific and clinical attention, the burden concept has attracted little in the way of theoretical attention. This has led to the absence of definitional consensus, which has, in turn, complicated the effort to provide ontological support for burden-related research. The present paper seeks to address these gaps by introducing the Biomedical Burden Ontology (BBO), a formal framework designed to represent and integrate burden-related data within biomedical informatics. The BBO is grounded in the Atlassian view of burden, which conceptualises burden as an individual’s obligatory participation in non-preferred processes. The ontology is implemented in the Web Ontology Language (OWL) and leverages Basic Formal Ontology (BFO), alongside existing ontologies such as the Mental Functioning Ontology (MFO) and the Emotion Ontology (MFOEM). Additionally, the BBO incorporates insights from predictive processing theories of brain function, framing burden as a disruption of an individual’s capacity to fulfil ‘optimistic’ predictions. By providing a structured approach to representing burden, the BBO facilitates research into patient experience, supports the development of minimally disruptive medicine, and enables more effective measurement of burden in clinical and policy contexts.

Mandevillian intelligence is a specific form of collective intelligence in which individual cognitive shortcomings, limitations, and biases play a positive functional role in yielding various forms of collective cognitive success. When this idea is transposed to the epistemological domain, mandevillian intelligence emerges as the idea that individual forms of intellectual vice may, on occasion, support the epistemic performance of some form of multi-agent ensemble, such as a socio-epistemic system, a collective doxastic agent, or an epistemic group agent. As a specific form of collective intelligence, mandevillian intelligence is relevant to a number of debates in social epistemology, especially those that seek to understand how group (or collective) knowledge arises from the interactions between a collection of individual epistemic agents.

Artificial General Intelligence (AGI) is often conceived as a self-contained system whose generality derives from increasingly powerful internal architectures. This paper challenges that assumption by developing an alternative, extended conception of AGI inspired by theories of active externalism and the extended mind. On this view, general intelligence is grounded not primarily in internal computation, but in a system’s capacity to construct and exploit extended mechanisms that incorporate environmental resources into its cognitive routines. We illustrate this idea with reference to the capabilities of large language models (LLMs), focusing on linguistic self-stimulation, tool use, and the creation of ad hoc computational tools. Across all these cases, language plays a peculiarly important role in enabling LLMs to leverage extended mechanisms that augment their cognitive and computational performances.

The mind–technology problem refers to issues that lie at the intersection of technology development and the philosophy of mind. In the present paper, I explore one aspect of the mind–technology problem, namely, the role of technologies in supporting the emergence of extended minds. I approach this issue from an engineering perspective, suggesting that the project to build extended minds yields insights into a number of philosophical problems. These include our understanding of the criteria for cognitive extension and the way the borders/boundaries of extended cognitive mechanisms are delineated.

Proponents of the extended mind insist that human states and cognitive processes can, at times, include non-biological resources that lie external to the bodily boundaries. In the present chapter, we apply this idea to large language models (LLMs), suggesting that some LLMs exist as extended cognitive (or computational) systems. We focus in particular on LLMs that exploit retrieval-augmented generation (RAG) techniques and online computational tools, proposing that these systems constitute extended architectures whose capabilities are realized, in part, by external structures. Drawing on parallels with classic cases of human cognitive extension, we argue that extended LLMs raise important philosophical questions concerning the nature and limits of artificial intelligence. We also draw attention to the role of language in supporting the emergence of extended systems. LLMs, it is suggested, are particularly well placed to benefit from their immersion in an online ecology that is home to a sizable chunk of human symbolic and technological culture.

Recent research has relied on the use of fine-tuning techniques to incorporate philosophical knowledge into Large Language Models (LLMs). The present paper outlines an alternative approach to the development of such systems—one that is rooted in a technique known as Retrieval-Augmented Generation (RAG). In contrast to fine-tuning, RAG does not seek to adjust the internal parameters (or internal memory) of an LLM. Instead, RAG relies on the retrieval of information from an externally-situated store, which functions as a form of non-parametric (or external) memory. Applying this technique to the works of the contemporary philosopher Andy Clark yields Digital Andy: an LLM that is able to respond to questions about the extended mind. This serves as a practical demonstration of RAG-based techniques, highlighting how philosophical knowledge can be ‘incorporated’ into an LLM without the need for additional machine learning. But Digital Andy’s reliance on extra-systemic resources also raises questions about the scope of active externalist theorizing, encouraging us to consider Digital Andy’s status as an extended cognitive/computational system. Addressing these questions reveals some interesting points of convergence between the philosophical effort to understand the extended mind and the technological effort to build the next generation of LLMs.

This chapter explores the cognitive and epistemic implications of emerging digital technologies from the standpoint of two philosophical positions: active externalism and virtue reliabilism. Emerging digital technologies (especially those associated with the Internet and World Wide Web) are important because they help to highlight issues that are not so easily revealed by the rather mundane and technologically low-grade examples favored by the philosophical community. The present analysis suggests that the general thrust of technology design is largely consistent with the criteria that have been used to evaluate putative cases of cognitive extension. In addition, the present analysis suggests that active externalism and virtue reliabilism are broadly compatible when it comes to the notion of extended knowledge. Despite this, a consideration of both the properties of emerging digital technologies and the requirements for extended knowledge reveals something of an unexpected tension between our prospective status as extended cognizers and extended knowers.

The terms "extended cognition" and the "extended mind" identify two strands of philosophical argument that are commonly subsumed under the general heading of active externalism. The present paper describes an integrated approach to understanding extended cognition and the extended mind—one that papers over the differences between these two, ostensibly distinct, forms of cognitive extension. As an added bonus, the paper describes how active externalism might be applied to the realm of non-cognitive phenomena, thereby yielding an expansion in the theoretical and empirical scope of the active externalist enterprise. Both these points of progress stem from what is called the dispositional hypothesis. According to the dispositional hypothesis, extended cognition occurs when the mechanisms responsible for the manifestation of dispositional properties include components that lie beyond the borders of the thing to which the dispositional properties are ascribed.

The present chapter discusses the implications of virtual reality for the theory and practice of embodied cognitive science. The chapter discusses how recent technological innovations are poised to reshape our understanding of the materially-embodied and environmentally-situated mind, providing us with a new means of studying the mechanisms responsible for intelligent behavior. The chapter also discusses how a synthetically-oriented shift in our approach to embodied intelligence alters our view of familiar problems, most notably the distinction between embedded and extended cognition. The chapter concludes by challenging the idea that virtual worlds ought to be seen as nothing more than a scientific ‘sandpit’ for the study of the embodied mind. Some forms of embodied intelligence can, I suggest, exist solely in virtual reality. In this sense, the ‘Matrix’ is not just a means of testing ideas about the forms of embodied intelligence that inhabit our own physical reality; it is also a means of creating entirely new forms of embodied intelligence. It is by building virtual worlds that we open the door to a new reality—one that is home to a potentially infinite variety of embodied, embedded, and perhaps even extended beings.

According to proponents of the extended mind, bio-external resources, such as a notebook or a smartphone, are candidate parts of the cognitive and mental machinery that realises cognitive states and processes. The present chapter discusses three areas of ethical concern associated with the extended mind, namely mental privacy, mental manipulation, and agency. We also examine the ethics of the extended mind from the standpoint of three general normative frameworks, namely, consequentialism, deontology, and virtue ethics.

This chapter explores the notion of the Web‐extended mind, which is the idea that the technological and informational elements of the Web can sometimes serve as part of the mechanistic substrate that realizes human mental states and processes. It is argued that while current forms of the Web may not be particularly suited to the realization of Web‐extended minds, new forms of user interaction technology as well as new approaches to information representation do provide promising new opportunities for Web‐based forms of cognitive extension. In addition, it is suggested that extended cognitive systems often rely on the emergence of social practices and conventions that shape how a technology is used. Web‐extended minds may thus depend on forms of socio‐technical co‐evolution in which social forces and factors play just as important a role as do the processes of technology design and development.

AI systems play an increasingly important role in shaping and regulating the lives of millions of human beings across the world. Calls for greater _transparency_ from such systems have been widespread. However, there is considerable ambiguity concerning what “transparency” actually means, and therefore, what greater transparency might entail. While, according to some debates, transparency requires _seeing through_ the artefact or device, widespread calls for transparency imply _seeing into_ different aspects of AI systems. These two notions are in apparent tension with each other, and they are present in two lively but largely disconnected debates. In this paper, we aim to further analyse what these calls for transparency entail, and in so doing, clarify the sorts of transparency that we should want from AI systems. We do so by offering a taxonomy that classifies different notions of transparency. After a careful exploration of the different varieties of transparency, we show how this taxonomy can help us to navigate various domains of human–technology interactions, and more usefully discuss the relationship between technological transparency and human agency. We conclude by arguing that all of these different notions of transparency should be taken into account when designing more ethically adequate AI systems.

Examples of extended cognition typically involve the use of technologically low-grade bio-external resources (e.g., the use of pen and paper to solve long multiplication problems). The present paper describes a putative case of extended cognizing based around a technologically advanced mixed reality device, namely, the Microsoft HoloLens. The case is evaluated from the standpoint of a mechanistic perspective. In particular, it is suggested that a combination of organismic (e.g., the human individual) and extra-organismic (e.g., the HoloLens) resources form part of a common mechanism that realizes a bona fide cognitive routine. In addition to demonstrating how the theoretical resources of neo-mechanical philosophy might be used to evaluate extended cognitive systems, the present paper illustrates one of the ways in which mixed reality devices, virtual objects (i.e., holograms), and online (Internet-accessible) computational routines might be incorporated into human cognitive processes. This, it is suggested, speaks to the recent interest in mixed/virtual reality technologies across a number of disciplines. It also introduces us to issues that cross-cut disparate fields of philosophical research, such as the philosophy of science and the philosophy of technology.

Trustworthiness is typically regarded as a desirable feature of national identification systems (NISs); but the variegated nature of the trustor communities associated with such systems makes it difficult to see how a single system could be equally trustworthy to all actual and potential trustors. This worry is accentuated by common theoretical accounts of trustworthiness. According to such accounts, trustworthiness is relativized to particular individuals and particular areas of activity, such that one can be trustworthy with regard to some individuals in respect of certain matters, but not trustworthy with regard to all trustors in respect of every matter. The present article challenges this relativistic approach to trustworthiness by outlining a new account of trustworthiness, dubbed the expectation-oriented account. This account allows for the possibility of an absolutist (or one-place) approach to trustworthiness. Such an account, we suggest, is the approach that best supports the effort to develop NISs. To be trustworthy, we suggest, is to minimize the error associated with trustor expectations in situations of social dependency (commonly referred to as trust situations), and to be trustworthy in an absolute sense is to assign equal value to all expectation-related errors in all trust situations. In addition to outlining the features of the expectation-oriented account, we describe some of the implications of this account for the design, development, and management of trustworthy NISs.

Proponents of the extended mind have suggested that phenomenal transparency may be important to the way we evaluate putative cases of cognitive extension. In particular, it has been suggested that in order for a bio-external resource to count as part of the machinery of the mind, it must qualify as a form of transparent equipment or transparent technology. The present paper challenges this claim. It also challenges the idea that phenomenological properties can be used to settle disputes regarding the constitutional status of bio-external resources in episodes of extended cognizing. Rather than regard phenomenal transparency as a criterion for cognitive extension, we suggest that transparency is a feature of situations that support the ascription of certain cognitive/mental dispositional properties to both ourselves and others. By directing attention to the forces and factors that motivate disposition ascriptions, we arrive at a clearer picture of the role of transparency in arguments for extended cognition and the extended mind. As it turns out, transparency is neither necessary nor sufficient for cognitive extension, but this does not mean that it is entirely irrelevant to our understanding of the circumstances in which episodes of extended cognizing are apt to arise.

An emerging consensus in cognitive science views the biological brain as a hierarchically-organized predictive processing system that relies on generative models to predict the structure of sensory information. Such a view resonates with a body of work in machine learning that has explored the problem-solving capabilities of hierarchically-organized, multi-layer neural networks, many of which acquire and deploy generative models of their training data. The present chapter explores the extent to which the ostensible convergence on a common neurocomputational architecture might provide inroads into the problem of digital immortality. In contrast to approaches that seek to recapitulate the physical structure of the human brain, the present chapter advocates an approach that is rooted in the use of machine learning algorithms. The claim is that a future form of deep learning system could be used to acquire generative models of a given individual or the sensory data that is processed by the brain of a given individual during the course of their biological life. The differences between these two forms of digital immortality are explored, as are some of the options for digital resurrection.

There have been a number of attempts to apply mechanism-related concepts to the notion of extended cognition. Such accounts appeal to the idea that extended cognitive routines are realized by mechanisms that transcend some salient border or boundary. The present paper describes some of the challenges confronting the effort to develop a mechanistic account of extended cognition. In particular, it describes five problems that must be resolved if we are to make sense of the idea that extended cognition can be understood via an appeal to extended mechanisms. These problems are intended to guide the philosophical effort to develop a mechanistic account of extended cognition.

The application of extended mind theory to the Internet and Web yields the possibility of Internet-extended knowledge—a form of extended knowledge that arises as a result of an individual's interactions with the online environment. The present paper seeks to advance our understanding of Internet-extended knowledge by describing the functionality of a real-world application, called the HoloArt app. In part, the goal of the paper is illustrative: it is intended to show how recent advances in mixed reality, cloud-computing, and machine intelligence might be combined so as to yield a putative case of Internet-extended knowledge. Beyond this, however, the paper is intended to support the philosophical effort to understand the notions of extended knowledge and the extended mind. In particular, the HoloArt app raises questions about the universality of some of the criteria that have been used to evaluate putative cases of cognitive extension. The upshot is a better appreciation of the way in which claims about extended knowledge and the extended mind might be affected by a consideration of technologically-advanced resources.

Arguments for extended cognition and the extended mind are typically directed at human-centred forms of cognitive extension—forms of cognitive extension in which the cognitive/mental states/processes of a given human individual are subject to a form of extended or wide realization. The same is true of debates and discussions pertaining to the possibility of Web-extended minds and Internet-based forms of cognitive extension. In this case, the focus of attention concerns the extent to which the informational and technological elements of the online environment form part of the machinery of the (individual) human mind. In this paper, we direct attention to a somewhat different form of cognitive extension. In particular, we suggest that the Web allows human individuals to be incorporated into the computational/cognitive routines of online systems. These forms of computational/cognitive extension highlight the potential of the Web and Internet to support bidirectional forms of computational/cognitive incorporation. The analysis of such bidirectional forms of incorporation broadens the scope of philosophical debates in this area, with potentially important implications for our understanding of the foundational notions of extended cognition and the extended mind.

Social machines are a prominent focus of attention for those who work in the field of Web and Internet science. Although a number of online systems have been described as social machines, there is, as yet, little consensus as to the precise meaning of the term “social machine.” This presents a problem for the scientific study of social machines, especially when it comes to the provision of a theoretical framework that directs, informs, and explicates the scientific and engineering activities of the social machine community. The present paper outlines an approach to understanding social machines that draws on recent work in the philosophy of science, especially work in so-called mechanical philosophy. This is what might be called a mechanistic view of social machines. According to this view, social machines are systems whose phenomena are explained via an appeal to socio-technical mechanisms. We show how this account is able to accommodate a number of existing attempts to define the social machine concept, thereby yielding an important opportunity for theoretical integration.

Arguments for the extended mind suggest the possibility of extended knowers, individuals whose epistemic standing is tied to the operation of cognitive circuits that extend beyond the bounds of skin and skull. When applied to the Internet, this idea yields the possibility of Internet-extended knowledge, a form of extended knowledge that derives from our interactions and engagements with the online environment. This, however, yields a tension: proponents of the extended mind have suggested that cognitive extension requires the automatic endorsement of bio-external content, but such a commitment appears to conflict with the need to carefully evaluate the veracity of online information. Lukas Schwengerer proposes a solution to this problem. He suggests that proponents of the extended mind ought to accept an alternative conception of the extended mind, one that allows for the evaluation of online content via the exercise of intellectual virtue. We highlight a number of problems with this proposal and suggest that further research is required before a virtue-theoretic approach to Internet-extended knowledge can be adopted. On a more positive note, we suggest that the exercise of intellectual virtue may play an important role in the developmental emergence of Internet-extended knowers.

The World Wide Web (Web) provides access to a global space of information assets and computational services. It also, however, serves as a platform for social interaction (e.g., Facebook) and participatory involvement in all manner of online tasks and activities (e.g., Wikipedia). There is a sense, therefore, that the advent of the Social Web has transformed our understanding of the Web. In addition to viewing the Web as a form of information repository, we are now able to view the Web in more social terms. In particular, it has become possible to see the Web as providing access to the human social environment. This is important, because issues of social embedding and social interaction have been seen to contribute to the emergence of human cognitive capabilities. The ability of the Web to provide access to the human social environment thus raises an important question: Can humanity play a productive role in the emergence of advanced forms of machine intelligence by virtue of their interactions and engagements with the online realm? The present paper attempts to show why this question is worth asking. It also attempts to highlight some of the ways in which Web-based forms of contact with the human social environment may be relevant to research into machine intelligence.

The World Wide Web has had a notable impact on a variety of epistemically-relevant activities, many of which lie at the heart of the discipline of knowledge engineering. Systems like Wikipedia, for example, have altered our views regarding the acquisition of knowledge, while citizen science systems such as Galaxy Zoo have arguably transformed our approach to knowledge discovery. Other Web-based systems have highlighted the ways in which the human social environment can be used to support the development of intelligent systems, either by contributing to the provision of epistemic resources or by helping to shape the profile of machine learning. In the present paper, such systems are referred to as ‘knowledge machines’. In addition to providing an overview of the knowledge machine concept, the present paper reviews a number of issues that are associated with the scientific and philosophical study of knowledge machines. These include the potential impact of knowledge machines on the theory and practice of knowledge engineering, the role of social participation in the realization of intelligent systems, and the role of standardized, semantically-enriched data formats in supporting the ad hoc assembly of special-purpose knowledge systems and knowledge processing pipelines.

The Internet provides access to a global space of information assets and computational services. It also, however, serves as a platform for social interaction (e.g., Facebook) and participatory involvement in all manner of online tasks and activities (e.g., Wikipedia). There is a sense, therefore, that the Internet yields an unprecedented form of access to the human social environment: it provides insight into the dynamics of human behavior (both individual and collective), and it additionally provides access to the digital products of human cognitive labor (again, both individual and collective). Such access is interesting from the standpoint of research into machine intelligence, for the human social environment looks to be of crucial importance when it comes to the evolutionary and developmental origins of the human mind. In the present paper, we develop a theoretical account that sees the Internet as providing opportunities for online systems to function as socially-situated agents. The result is a vision of machine intelligence in which advanced forms of cognitive competence are seen to arise from the creation of a new kind of digital socio-ecological niche. The present paper attempts to detail this vision with respect to the notion of socially-scaffolded cognition. It also describes some of the forms of machine learning that may be required to enable online systems to press maximal cognitive benefit from their new-found informational contact with the human social world.