This research reviews explanation and interpretation for Explainable Artificial Intelligence (XAI) methods in order to boost complex machine learning model interpretability. The study shows the influence and belief of XAI in users that trust an Artificial Intelligence system and investigates ethical concerns, particularly fairness and biasedness of all the nontransparent models. It discusses the shortfalls related to XAI techniques, putting crucial emphasis on extended scope, enhancement and scalability potential. A number of outstanding issuesespecially in need of further work can (...) involve standardization, user-centered design and interdisciplinary in strategies for improving the practical utility of XAI. (shrink)

A dominant claim in the philosophical literature on trust is that we should stop thinking in terms of group trustworthiness or appropriate trust in groups. In this paper we push back against this claim by arguing that philosophical work on trust would benefit from being brought into closer contact with empirical work on the nature of trust. We consider data on reactive attitudes and moral responsibility to adjudicate on different positions in the philosophical literature on (...) class='Hi'>trust. An implication of our argument is that the distinction between different kinds of groups – mere groups versus institutional groups – deserves more attention than is currently recognized in the philosophical literature on trust. (shrink)

This commentary offers a coherent reading of the papers presented in the special issue ‘Exclusion, Engagement, and Empathy: Reflections on Public Participation in Medicine and Technology’. Focusing on intellectual self-trust it adds a further perspective on the harmful epistemic consequences of social exclusion for individual agents in healthcare contexts. In addition to some clarifications regarding the concepts of ‘intellectual self-trust’ and ‘social exclusion’ the commentary also examines in what ways empathy, engagement and participatory sense-making could help to avoid (...) threats to intellectual self-trust that arise form being excluded from participation in communicative practices in the context of healthcare. (shrink)

Human relationships have been explained through various elements such as emotion, social norms, and roles. However, these approaches have limitations in providing an integrated explanation of the process through which relationships are formed, stabilized, and transformed over time. In particular, emotion-centered explanations are effective in explaining the initial trigger of relationship formation but fail to sufficiently account for persistence and stability, while norm- or role-based approaches can present the external framework of relationships but are limited in explaining how structures are (...) formed within actual interaction. Based on these limitations, this paper understands human relationships as a dynamic structure in which emotion, ethical interaction, and trust are connected over time. Emotion functions as the starting point of relationship formation by triggering interaction, and through repetitive interaction, behavioral patterns such as responsibility, consistency, and respect for boundaries are formed, which constitute the structure of relationships. As this process accumulates, predictability and mutual expectations emerge between individuals, leading to the formation of trust. Furthermore, this paper argues that emotion, ethical interaction, and trust do not proceed in a one-directional sequence, but rather form a cyclical structure in which established trust reshapes subsequent emotional experiences and patterns of interaction. This cyclical structure functions as an integrated theoretical framework for explaining how human relationships are formed, stabilized, and weakened over time. Accordingly, this paper does not reduce human relationships to a single element, but presents an analytical framework that explains, in an integrated manner, the starting point of relationship formation, the formation of structure through repetitive interaction, and the cyclical process of change centered on trust. (shrink)

The doxastic view (DV) of trust maintains that trust essentially involves belief. In a recent paper, Arnon Keren (Citation2020) gestures toward a new objection to the view, labeled Trust’s Meno Problem (TMP), which calls into question the DV’s ability to explain the widely held intuition that trust has distinct and indispensable value. As of yet, there has been no attempt to take up TMP on behalf of DV. This paper aims to fill precisely this lacuna. I (...) do so in three main stages. In §1 I contextualize and elucidate the problem, to which Keren gestures but does not address in detail. In §2 I disambiguate multiple possible interpretations of TMP, seeking to identify the most philosophically challenging. Finally, in §3, I argue that DV can solve even this interpretation. In order to do so, I make use ofthe highly plausible claim we find in the work of Katherine Hawley (Citation2012, 2019): that trust pays a compliment to the trustee. The payoffs of exploring the doxastic view in the context of Trust’s Meno Problem are twofold: we better understand the nature of the problem itself, and we see that the doxastic view can give a satisfying answer. (shrink)

Human relationships are formed and maintained through interaction, yet it is consistently observed that even under identical interactional conditions, the outcomes of relationships—whether maintained, transformed, or broken—differ across individuals. Existing theories of relationships have primarily explained this phenomenon through emotional factors or interactional structures, but they have failed to sufficiently account for why the same structure produces different relational outcomes. To address this problem, this paper focuses on the idea that the outcomes of relationships are determined not by the interaction (...) structure itself, but by the internal conditions through which that structure is interpreted and processed by individuals. To this end, the paper distinguishes recognition from respect and analyzes respect as a non-automatic process that does not necessarily follow from recognition. It further introduces respect capacity as the ability to maintain the boundaries and rights of others, and systematically examines its components and variability. In addition, the paper formalizes the concept of expectation distortion by showing that expectations can be formed prior to interaction through projection processes and can be maintained and reinforced through cognitive distortion. Trust is defined not as an emotional state, but as the accumulated result of repeated choices, and is analyzed as a structure formed around predictability and the willingness to maintain the relationship, which can gradually collapse over time. Based on these discussions, the paper presents a sequential structure of recognition–respect–interaction–expectation–trust, emphasizing that this entire process is filtered through the internal conditions of individuals. Through this, it explains the mechanism by which relational outcomes diverge under identical interactional structures and proposes a theoretical framework for analyzing the formation, maintenance, and breakdown of human relationships at the level of individual differences. (shrink)

Over the past couple of decades, there has been growing concern about the alleged rise of various forms of science denial. But what exactly is science denial? How does it differ from ordinary scientific ignorance? Is it really on the rise? If so, what explain this trend? And what is so concerning about it in the first place? This Element has four goals. Its first (and least ambitious) goal is to is to bring some conceptual clarity by developing a clearer (...) notion of science denial and gaining a better understanding of the concerns raised by the phenomenon it is meant to capture. Its second (and more ambitious) goal is to distinguish among several specific phenomena that are often conflated under the label ‘science denial.’ Its third (and even more ambitious) goal is to argue that none of these specific phenomena warrants all of the concerns that motivate its critics. Its fourth (and most ambitious) goal is to expose the inadequacy of a popular diagnosis of the epistemic malaise afflicting liberal democracies – the post-truth diagnosis – and to sketch an alternative to it – the post-trust diagnosis. (shrink)

While the concerns about media trust have historical roots, the dynamics have evolved with technological advancements, the rise of digital media, and the challenges posed by disinformation in the 21st century. Deepfake – the media content (text, images, videos) created by AI technology – is just another concern for journalists, politicians and society in the current media trust crisis. It is a tool that can be successfully applied in an information warfare. To be honest, face and voice spoofing (...) has always existed, but deepfake technology may now make fake videos faster, cheaper, with high quality and is accessible to anybody. Although deepfake videos have the potential to undermine trust in journalism, their actual impact on audiences remains limited due to information overload, increased suspicion and skeptical attitude, decreased trust in media and advanced audience adaptability for orientation in the information flow. (shrink)

The point of departure of this paper is the striking similarities between the dispositions critical media literacy education aims to cultivate and the characteristics conspiracy theorists claim to embody. The golden question of critical literacy, “who benefits?” is in fact the central question of conspiracy theorists: “cui bono?” While critical media literacy educators teach learners to disrupt the common sense, to interrogate multiple viewpoints, to focus on sociopolitical issues, and to take actions and promote social justice, conspiracy theorists claim that (...) they do exactly those things. What I wish to illuminate through this juxtaposition, however, is not a way to clearly demarcate “critical” from “conspiratorial” theorizing, but is a question of educational desirability of skepticism. Skepticism in this paper is not understood as a classical epistemological question of the possibility of knowledge attainment, but as a form of vigilance toward epistemic authorities like the government, media, and academic institutions. Drawing on philosophy of conspiracy theories and political epistemology, I illuminate the problem of radical skepticism that characterizes some conspiracy theories and discuss the pivotal epistemic role of trust. Through this, I claim that critical media literacy education for democratic citizenship must strike a balance between cultivation of skepticism and building of trust toward epistemic authorities. (shrink)

An assurance theory of trust begins from the act of assurance – whether testimonial, advisorial or promissory – and explains trust as a cognate stance of resting assured. My version emphasizes the risks and rewards of trust. On trust’s rewards, I show how an assurance can give a reason to the addressee through a twofold exercise of ‘normative powers’: (i) the speaker thereby incurs an obligation to be sincere; (ii) if the speaker is trustworthy, she thereby (...) gives her addressee the reason. Each claim is controversial; my contribution lies in how I defend them together, linking the sincerity obligation to the provision of reasons. On trust’s risks, I defend a popular thesis about the nature of trust – that trust differs from mere reliance in how it risks betrayal – against the objection that the thesis ‘moralizes’ trust. Such betrayal need not be moral, I argue, because it may derive entirely from how the assurance purports to provide a reason. (shrink)

Philosophers have had much to say about the moral obligations arising from trust. These obligations involve, predominantly, interpersonal relationships. But what can we say about the moral relevance of trust in institutional settings? In this paper, we consider the particularised approach to trust and its focus on interpersonal relationships and argue that it is far from clear whether this is the kind of relationship that persons can have with respect to firms. For a particularised view of (...) class='Hi'>trust to be applicable to firms, firms must be seen as institutions that are to some extent distinct from their individual members. They would also have to be fitting targets and sources of reactive attitudes in morally relevant trust relationships. Such an account of trust in firms is not as yet forthcoming. To inspire future work on trust and firms, we take note of an argumentative strategy in the literature on corporate agency that lets us conceive of firms as appropriate targets and/or sources of reactive attitudes. In so far as this strategy turns out to have empirical efficacy, we suggest it may also prove fruitful for thinking about trust in relation to firms. (shrink)

Throughout their history, methodologies like Design for Values have taken trust as an important value in relation to technology. At first, this referred exclusively to the idea that technology influences human trust relationships, and that technology developers should take interpersonal trust into consideration as a value to be optimised and balanced. More recently, it has become common in policy and technology design circles to hold that technologies themselves are possible objects of trust and distrust, and that (...) autonomous technologies such as AI should be designed to be trustworthy. However, the value of trust in technology can be problematic in various ways: some philosophers have warned that trusting technology is impossible, misleading, or even dangerous. In this chapter, after an introduction of conceptions of trust, we set out a moderate view on which trustworthiness can reasonably be attributed to technology. Making technologies trustworthy has limitations because trustworthiness and trust are only loosely related, and because trust is not an unconditional good. We examine three core examples here, namely trust in the context of social media platforms, neurotechnology and “trustworthy” generative AI. Looking at these examples, we argue for a rich conception of trust in, and through, technology, that is open to interdisciplinary treatment, and thus open for empirical and technical exploration. (shrink)

This paper presents an exhaustive historical and mathematical survey of the seventeen Permanent Axioms underlying the first machine-verified Coq formalisation of a global regularity result for the three-dimensional incompressible Navier-Stokes equations on the periodic torus T³. The formalisation establishes subcritical energy estimates for arbitrary smooth initial data in the Sobolev regularity class H^s for s > 3, producing a combined energy inequality that precludes finite-time blowup via a Gronwall-Bihari closure. The proof chain comprises approximately 59 Qed theorems, zero Admitted statements, (...) and 17 Permanent Axioms encoding classical results of PDE theory and harmonic analysis spanning 1919 to 1989. Each axiom is traced to its original publication, its proof history, its modern textbook treatments, its validity in the specific setting of incompressible flow on T³, its relationship to the broader Navier-Stokes regularity literature, and its current formalisation status in proof assistants including Lean, Coq, and Isabelle. The axioms encode results by Leray (1934), Sobolev (1938), Calderón and Zygmund (1952), Gronwall (1919), Kato and Ponce (1988), Prodi (1959), Serrin (1963), Giga (1981), Beale, Kato, and Majda (1984), Batchelor (1967), Plancherel (1910), Bernstein (1912), and Stein (1970). The paper introduces the concept of the Permanent Axiom Trust Boundary: the epistemic distinction between an author's novel derivations (which are entirely machine-verified) and the classical canon of twentieth-century mathematics (which the formalisation assumes as axiomatic). This boundary defines what machine verification can and cannot certify, and constitutes a contribution to the epistemology of computer-assisted mathematical proof. The formalisation gap between classical continuous mathematics and machine-verified type theory is documented, with a projected closure timeline informed by the HALF Project at Bonn, the Carleson formalisation, and the Armstrong-Kempe De Giorgi-Nash-Moser programme. All seventeen axioms are confirmed as Classical and Standard. Zero require further justification. The trust required by the formalisation is not trust in the author but trust in the established canon of analysis and PDE theory. (shrink)

The correct identification of the abuse of political power during the COVID-19 crisis remains a challenge because officially declaring the pandemic allowed political representatives to exercise additional power disguisable as the maintenance of functioning social order under the principle of preserving humankind. One way to observe the abuse of power in its excess is the degree of compliance exhibited by the people who laid juridical restrictions for the purpose of combating COVID-19. The behaviour of political representatives was evidence of political (...) exceptionalism, for political representatives, as decisive authorities, were the first to disregard measures introduced without suffering negative consequences. Such abuse catalysed the collapse of trust in state communities, but has for its true result the consolidation of the class system in which the ethical value of trust is played out. In this paper, I discuss how the behaviour of state representatives demonstrates the existence of political exceptionalism based on the class system and what this means for the persistence of national trust. (shrink)

At least since W. V. O. Quine's famous critique of the analytic/synthetic distinction, philosophers have been deeply divided over whether there are any analytic truths. One line of thought suggests that the simple fact that people have ' intuitions of analyticity' might provide an independent argument for analyticities. If defenders of analyticity can explain these intuitions and opponents cannot, then perhaps there are analyticities after all. We argue that opponents of analyticity have some unexpected resources for explaining these intuitions and (...) that, accordingly, the argument from intuition fails. (shrink)

This report documents an independent verification exercise on the Permanent Axiom trust boundary of a machine-verified Coq formalisation in the analysis of partial differential equations. The underlying formalisation consists of three independent Coq files for subcritical energy estimates of 3D incompressible Navier-Stokes on the periodic torus, verified by the Coq kernel with zero Admitted statements. The question the report addresses is narrower and more methodological than the mathematics itself: when a formal proof is accepted by the kernel, what remains (...) to be verified by a human auditor, and how should that residual verification work be structured? The report's methodological contribution is an explicit taxonomy of the hypothesis-to-axiom pipeline in AI-assisted formal mathematics. Every Hypothesis introduced in the formalisation is classified as either (a) discharged by a corresponding Lemma whose statement is literally identical, or (b) declared as a Permanent Axiom with a documented classical reference, with no third category permitted. The trust boundary is the set of Permanent Axioms — seventeen in total — each of which is cited to a published classical result, and each of which is preserved in the audit history rather than silently inlined into the proof chain. The report argues that this explicit classification is the minimum honest standard for any independent-researcher formal verification project, and that the alternative — quietly accepting hypotheses without distinguishing between proved and axiomatised ones — produces a formalisation whose trust structure is unreadable. The report is offered as a contribution to the philosophy of mathematical practice and to the methodology of AI-assisted formal verification. The author is a historian by training. The report accompanies a full Figshare deposit containing the Coq files, the manuscript, and supporting documents, and is part of the broader FIELDS Protocol for AI-assisted machine-checked mathematics. (shrink)

For nearly seven decades, psychology has operated upon a mistaken ontology of truth. What Happens When You Clap restores its philosophical foundation by showing that knowing is relational, fiduciary, and co-authored—not alignment with an external constant. -/- This work offers a landmark synthesis bridging psychology, philosophy, and governance. It redefines cognition, trust, and truth as one fiduciary process linking the dynamics of individual minds with the ethics of institutions. The inquiry begins with a simple scene: a theatre audience applauding (...) a mediocre performance, and one spectator who hesitates to join. From this moment unfolds the Relational Theory of Epistemic Clientelism (RTEC)—an extension of Epistemic Clientelism Theory (ECT)—which explains how truth is continually negotiated through trust, recognition, and dependence. The hesitation before applause becomes a micro-experiment in epistemic life, revealing how knowledge arises not from correspondence with an external reality but from fiduciary relations between perceiving minds. -/- Re-examining the classic experiments of Asch, Festinger, and Milgram, the study challenges their assumption of fixed reality and reframes dissonance as an epistemic event—the lived signal that knowing is relational and contingent. Drawing on phenomenology, social psychology, fiduciary ethics, and political theory, it argues that truth endures only through candour and collapses when performance replaces sincerity. Micro-acts of recognition, repeated and amplified, generate the macro-structures of authority that govern audiences, universities, corporations, and states. -/- The conclusion marks a new frontier: epistemic psychology as a unifying science of fiduciary cognition. It proposes fiduciary pluralism as the ethical architecture of modern societies—institutions as trustees of epistemic diversity, citizens as stewards of candour. The study speaks to psychologists, cognitive scientists, philosophers, sociologists, educators, and governance theorists seeking a parsimonious model of cognition and trust, as well as to those designing AI or policy frameworks for epistemic integrity and democratic resilience. (shrink)

Summary: Edward Lanphier and colleagues contend that human germline editing is an unethical technology because it could have unpredictable effects on future generations. In our view, such misgivings do not justify their proposed moratorium.

The rise of personalised medicine can be seen as an extension of individualism and as a threat to the common good.

Consider the following patient: a 40-year-old man who has had back pain that radiates down his left leg, on and off for 2 months. He performs his normal activities and does not have any “red flag”...

The aim of this study is to explore the effects of fake news on consumers’ brand trust in the food security context. The starting point of our research is the finding that issues related to food security cannot be addressed without the contribution of multinational food corporations. The efficiency of their intervention depends on their capacity to build and preserve their brand trust despite the multifarious fake news stories that contaminate the information flow. Is brand trust sensitive (...) to fake news? In some cases, the spread of fake news in mass media and social media negatively affects food companies. In other cases, consumers’ trust remains relatively unchanged. These ambivalent reactions give us good reason to assess whether consumers’ exposure to negative fake news influences their trust in international food brands. Using a one-group pretest-posttest research design, we found that the effects of fake news on consumers’ brand trust are predominantly negative, but in a few cases, these effects can be neutral or positive. These results could be useful for PR and marketing researchers and professionals interested in fake news phenomena and brand trust because they shed light on the real threat fake news represents for multinational food companies. (shrink)

Digital twins were built to close the gap between a model and the world it represents. An aircraft engine, a patient, a factory floor, each twin earns its keep by tracking reality more closely as data accumulates. Generative AI unsettles that premise. The same architecture, continuous modeling of a target, behavioral feedback, adaptive updating, can be pointed at a person instead of a machine, and optimized for sustained persuasion instead of predictive accuracy. We call this inversion the Digital Twin of (...) Deception (DTD): an adaptive computational model that tracks a target's beliefs, motivations, and trust state in order to shape their behavior while delaying detection. The existing literature linking digital twins to deception is almost entirely defensive, honeypots, decoy infrastructure, attacker modeling for industrial control systems. DTD names the adversarial mirror image that this literature does not address: cognitive modeling deployed against a person rather than on a person's behalf. We situate the concept against recent evidence on LLM-assisted romance fraud, deepfake-enabled executive impersonation, and the growing empirical record on AI persuasion, then propose the Trust Divergence Index (TDI), a measure of the gap between perceived credibility and independently verifiable evidence, as a way of detecting the pattern rather than any single deceptive claim. We close by asking what a certification regime for AI-mediated relationships would need that current AI-agent identity standards do not yet provide. (shrink)

According to most accounts of trust, you can only trust other people (or groups of people). To trust is to think that another has goodwill, or something to that effect. I sketch a different form of trust: the unquestioning attitude. What it is to trust, in this sense, is to settle one’s mind about something, to stop questioning it. To trust is to rely on a resource while suspending deliberation over its reliability. Trust (...) lowers the barrier of monitoring, challenging, checking, and questioning. Trust sets up open pipelines between yourself and parts of the external world. Trust permits external resources to have a similar relationship to one as one’s internal cognitive faculties. This creates efficiency, but at the price of exquisite vulnerability. We must trust in this way because we are cognitively limited beings in a cognitively overwhelming world. Crucially, we can hold the unquestioning attitude towards objects and technologies. When I trust my climbing rope, I stop worrying about its reliability. When I trust my online calendaring system, I simply go to the events indicated, without question. But, one might worry, how could one ever hold such a normatively loaded attitude as trust towards mere objects? How could it ever make sense to feel betrayed by an object? Trust is our engine for expanding and outsourcing our agency — for binding external processes into our practical selves. Thus, we can be betrayed by our smartphones in the same way that we can be betrayed by our memory. When we trust, we try to make something a part of our agency, and we are betrayed when our part lets us down. To unquestioningly trust something is to let it in—to attempt to bring it inside one’s practical functioning. This suggests a new form of gullibility: agential gullibility, which occurs when agents too hastily and carelessly integrate external resources into their own agency. (shrink)

The increasing use of artificial intelligence (AI) systems in our daily life through various applications, services, and products explains the significance of trust/distrust in AI from a user perspective. AI-driven systems have significantly diffused into various fields of our lives, serving as beneficial tools used by human agents. These systems are also evolving to act as co-assistants or semi-agents in specific domains, potentially influencing human thought, decision-making, and agency. Trust/distrust in AI plays the role of a regulator and (...) could significantly control the level of this diffusion, as trust can increase, and distrust may reduce the rate of adoption of AI. Recently, varieties of studies have paid attention to the variant dimension of trust/distrust in AI, and its relevant considerations. In this systematic literature review, after conceptualization of trust in the current AI literature review, we will investigate trust in different types of human-Machine interaction, and its impact on technology acceptance in different domains. In addition to that, we propose a taxonomy of technical (i.e., safety, accuracy, robustness) and non-technical axiological (i.e., ethical, legal, and mixed) trustworthiness metrics, and some trustworthy measurements. Moreover, we examine some major trust-breakers in AI (e.g., autonomy and dignity threat), and trust makers; and propose some future directions and probable solutions for the transition to a trustworthy AI. (shrink)

A rift has opened between philosophical and non-philosophical perspectives on what kind of conceptual stance we can take towards artificial intelligence: are AI-powered machines the sort of things that can be trusted? Outside of philosophy, an ever-increasing body of research is looking at features of AI that promote trust and trustworthiness to understand how and why people “trust” AI. Within philosophy, this entire endeavor is widely seen as resting on a conceptual confusion about trust; trust requires (...) the object of trust having agential capacities that modern AI lacks, and so whatever this body of research is studying, it is not, properly speaking, “trust”. Who is right? To shed light on this question, we explore the extent to which people thought sentences involving trusting different sorts of things make sense. We find that people do think it makes sense to talk about trusting AI, although less than it makes sense to trust other humans – even if they do not think AI has the sort of agential capacities philosophers think are required to be the object of trust. We conclude that either trust does not require agency or that there are multiple senses of trust, and neither conclusion is consistent with “trustworthy AI” being a conceptual confusion. While this undermines one criticism of AI and AI-related stakeholders, it nonetheless does not entail that AI is worthy of our trust. (shrink)

On extant accounts of trust, it is not apt to trust someone who seems or actually is untrustworthy. It also does not seem apt, on such accounts, to morally demand trust and blame people for not trusting. Yet, there seem to be cases where it is plausible that trusting the untrustworthy might be apt or where we might aptly demand trust and blame people for not trusting. In this paper, we sketch out an account of (...) class='Hi'>trust according to which to trust someone involves exercising a moral power which gives trustees permission or additional reasons to do as trusted. On this account, there can be moral reasons to trust other than the trustee’s reliability. This allows our account to better account for these cases than existing accounts of trust. (shrink)

According to doxastic accounts of trust, trusting a person to \(\varPhi \) involves, among other things, holding a belief about the trusted person: either the belief that the trusted person is trustworthy or the belief that she actually will \(\varPhi \) . In recent years, several philosophers have argued against doxastic accounts of trust. They have claimed that the phenomenology of trust suggests that rather than such a belief, trust involves some kind of non-doxastic mental attitude (...) towards the trusted person, or a non-doxastic disposition to rely upon her. This paper offers a new account of reasons for trust and employs the account to defend a doxastic account of trust. The paper argues that reasons for trust are preemptive reasons for action or belief. Thus the Razian concept of preemptive reasons, which arguably plays a key role in our understanding of relations of authority, is also central to our understanding of relations of trust. Furthermore, the paper argues that acceptance of a preemptive account of reasons for trust supports the adoption of a doxastic account of trust, for acceptance of such an account both neutralizes central objections to doxastic accounts of trust and provides independent reasons supporting a doxastic account. (shrink)

When are we obligated to believe someone? To what extent are people authorities about their own experiences? What kind of harm might we enact when we doubt? Questions like these lie at the heart of many debates in social and feminist epistemology, and they’re the driving issue behind a key conceptual framework in these debates—gaslighting. But while the concept of gaslighting has provided fruitful insight, it's also proven somewhat difficult to adjudicate, and seems prone to over-application. In what follows, I (...) argue that Katherine Hawley's theory of trust can provide a useful alternative lens for looking at contested testimony. To do this, I focus on a particularly complex—but increasingly popular—application of gaslighting: the physician/patient relationship, and the idea of ‘medical gaslighting’. I argue that, even though patients can experience harm when they are disbelieved, there are nevertheless good reasons for physicians not to trust patients about at least some of their own narratives. (shrink)

While trust is foundational to the doctor-patient relationship, the introduction of AI into healthcare settings poses the risk of eroding this trust, and such erosion cannot be countered simply by appealing to the notion of “trustworthy AI.” We argue that trust presupposes specific epistemic attitudes that cannot be meaningfully applied to AI systems. Accordingly, our focus is not on specifying which capabilities AI must exhibit in order to appear trustworthy, but on examining from an epistemological perspective how (...) the use of AI reshapes the dynamics of trust within the doctor-patient relationship. To this end, we first sketch conceptions of trust and demonstrate how trust differs from reliance. We then combine the model of Computational Reliabilism with an epistemic framework to develop a matrix for the ethical analysis of our use cases. Finally, we apply this framework to three scenarios of melanoma detection, risk prediction, and psychotherapy chatbots, which we construct by mapping epistemic stances across different modes of human-machine interaction, ranging from collaborative support with varying degrees of autonomy to the replacement of human-human interaction. We argue that the application of AI in the doctor-patient relationship exposes what we call a “reliability gap” — a conceptual space where the opaque nature of advanced AI systems prevents both doctors and patients from independently verifying their reliability. This creates a dynamic where reliability in the AI’s performance is increasingly mediated by the doctor as a proxy. Our use cases demonstrate that the more autonomous and opaque AI systems are, the more trust in the doctor becomes essential for bridging reliability gaps, while threatening to overburden the doctor’s central role. (shrink)

Trust not only disposes us to feel betrayed, trust can be betrayed. Understanding what a betrayal of trust is requires understanding how trust can ground an obligation on the part of the trusted person to act specifically as trusted. This essay argues that, since trust cannot ground an appropriate obligation where there is no prior obligation, a betrayal of trust should instead be conceived as the violation of a trust-based obligation to respect an (...) already existing obligation. Two forms of trust are evaluated for their potential to ground such a second-order obligation. One form counts as a gift to the trusted person only because it does not involve an expectation of trustworthiness; the other form cannot count as a gift but confers an honor because it does include an expectation of trustworthiness. Only trust that confers an honor generates a second-order obligation whose violation would be a betrayal of trust. -/- . (shrink)

In this paper, I argue that reactive feelings such as betrayal and personal disappointment are not inherent to the attitude of trust. Instead, such feelings are better understood as responses to impairments in relationships. Trust, I propose, is a fully doxastic mechanism that fundamentally consists of the belief that the trustee will follow through the norms constitutive of the relationship, such that a breach of trust directly calls only for an epistemic reassessment of the trustee’s trustworthiness. I (...) further show that what warrants reactive feelings is not the mere fact of trust being broken, but whether through the violation of trust, the person reveals a disregard for the relationship. (shrink)

Two topics at the center of Ethics of AI and HRI regard trust in AI agents as well as the adjudication of moral responsibility in situations where AI causes harm. In this paper we aim to advance the state of the art concerning these topics in several regards: First, we propose and evaluate a new empirical paradigm for measuring appropriate or calibrated trust in AI, that is, attitudes which are neither too trusting nor too cautious. The best way (...) to measure calibrated trust, we argue, is by contrasting trust vested in AI agents when their relevant capacities equal those of a human expert in the domain. A second shortcoming of extant work regards generalizability: Trust in, and reliance on, AI are standardly explored with respect to a single context or domain. To investigate context-sensitivity, we ran experiments (total N=1276) across five key areas of AI application. Finally, we explored perceived moral responsibility for harm caused in human-AI interaction, with a particular focus on recent philosophical debates on the topic. Our findings suggest that approximately half of the participants vest equal trust in AI and human agents when their capacities are the same. However, there is considerable variation in trust calibration across domains, suggesting that context-sensitivity needs more attention. Human agents are attributed more moral responsibility than AI agents, whereas their supervisors are blamed less than those of AI agents. This suggests that, at least according to folk morality, there are no perceived "responsibility gaps" (Matthias 2004; Sparrow 2007) and that "retribution gaps" are a genuine possibility (Danaher 2016). (shrink)

This paper sets out an account of trust in AI as a relationship between clinicians, AI applications, and AI practitioners in which AI is given discretionary authority over medical questions by clinicians. Compared to other accounts in recent literature, this account more adequately explains the normative commitments created by practitioners when inviting clinicians’ trust in AI. To avoid committing to an account of trust in AI applications themselves, I sketch a reductive view on which discretionary authority is (...) exercised by AI practitioners through the vehicle of an AI application. I conclude with four critical questions based on the discretionary account to determine if trust in particular AI applications is sound, and a brief discussion of the possibility that the main roles of the physician could be replaced by AI. (shrink)

This book offers a defence of Wrightean epistemic entitlement, one of the most prominent approaches to hinge epistemology. It also systematically explores the connections between virtue epistemology and hinge epistemology. -/- According to hinge epistemology, any human belief set is built within and upon a framework of pre-evidential propositions – hinges – that cannot be justified. Epistemic entitlement argues that we are entitled to trust our hinges. But there remains a problem. Entitlement is inherently unconstrained and arbitrary: We can (...) be entitled to any hinge proposition under the right circumstances. In this book, the author argues that we need a non-arbitrariness clause that protects entitlement from defeat. This clause, he argues, is to require epistemic virtue. Virtuous cognitive dispositions provide the non-arbitrariness clause that protects entitlement from defeat. The epistemic character of the agent who holds a particular set of hinges tells us something about the hinges’ epistemic status. Conversely, epistemic virtues are cognitive dispositions and capacities that rely on hinge propositions – without trusting in some hinges, we would be unable to exercise our virtues. -/- Trust Responsibly will appeal to scholars and advanced students working on epistemology, Wittgenstein, and virtues. (shrink)

The persistence of scientific misconceptions is often attributed to a decline in trust in experts. Against this simplistic picture, we emphasize that misplaced trust in expertise plays a crucial role in sustaining such misconceptions: even laypeople actively seeking expert guidance may nonetheless place their trust in unreliable sources. The paper identifies two main kinds of ’epistemic traps’ that are relevant to this phenomenon. In addition to fake experts who flaunt competence they lack (like pseudoexperts and pseudo-scientists), we (...) emphasise the importance of unreliable experts, who possess relevant credentials but systematically offer unreliable testimony. The resulting picture clashes with the commonplace idea that laypeople bear significant epistemic blame for endorsing misconceptions. Even responsible agents can be misled by unreliable experts who display legitimate credentials: the less evident the source’s unreliability, the less responsible is the agent for being misled. (shrink)

Where does entitlement to trust come from? When we trust someone to φ, do we need to have reason to trust them to φ or do we start out entitled to trust them to φ by default? Reductivists think that entitlement to trust always “reduces to” or is explained by the reasons that agents have to trust others. In contrast, anti-reductivists think that, in a broad range of circumstances, we just have entitlement to (...) class='Hi'>trust. even if we don’t have positive reasons to do so. In this paper, we argue for a version of anti-reductivism. Roughly, we argue that we have default entitlement to trust someone to φ so long as there is an operative norm that requires S to φ. At least in such circumstances (and absent defeaters), we don’t need any positive reasons to trust S to φ. (shrink)

When we learn that someone holds irrational beliefs, we often respond by reducing our epistemic trust in them. In this paper, I will propose a novel account of such trust reductions. The recently popular relationship-modification account (RMA) of epistemic blame will serve as a foil for this project. RMA says that epistemically blaming others for their epistemic failings involves modifying our epistemic relationships with them, paradigmatically via a reduction of epistemic trust. RMA has recently faced two challenges (...) of extensional inadequacy, which I will show result from a mistaken view about what type of response trust reductions are. I will draw on resources from legal theory to show that trust reductions bear all the hallmarks of so-called non-punitive measures, which serve preventative collective purposes. I will argue for an account of trust reductions as non-punitive epistemic measures that serve the purpose of preventing unreliable informants from negatively affecting the epistemic commons. My account explains when and why it is appropriate to reduce epistemic trust and shows where RMA goes wrong. (shrink)

I motivate and defend a new way of theorising about trust and trustworthiness – and their relationship to each other – by locating both within a broader picture that captures largely overlooked symmetries on both the trustor's and trustee's side of a cooperative exchange. The view defended here takes good cooperation as a theoretical starting point; on the view proposed, cooperation between trustor and trustee is working well when achievements in trust and in responding to trust are (...) matched on both sides of the trust exchange. In a bit more detail, the trustor ‘matches’ her achievement in trusting (an achievement in fitting reliance to reciprocity) with the trustee's achievement in responding to trust (an achievement in fitting reciprocity to reliance). From this starting point, we can then appreciate symmetrical ways that the trustor and trustee can (respectively) fall short, by violating what are shown to be symmetrical evaluative norms – of success, competence and aptness – that regulate the attempts made by both trustor and trustee. The overall picture has important advantages over the received way of theorising about how trust stands to trustworthiness, and it clears the way – by identifying key questions that have been obscured – to making further progress. (shrink)

First, we explain the conception of trustworthiness that we employ. We model trustworthiness as a relation among a trustor, a trustee, and a field of trust defined and delimited by its scope. In addition, both potential trustors and potential trustees are modeled as being more or less reliable in signaling either their willingness to trust or their willingness to prove trustworthy in various fields in relation to various other agents. Second, following Alfano (forthcoming) we argue that the social (...) scale of a potential trust relationship partly determines both explanatory and normative aspects of the relation. Most of the philosophical literature focuses on dyadic trust between a pair of agents (Baier 1986, Jones 1996, Jones 2012, McGeer 2008, Pettit 1995), but there are also small communities of trust (Alfano forthcoming) and trust in large institutions (Potter 2002, Govier 1997, Townley & Garfield 2013, Hardin 2002). The mechanisms that induce people to extend their trust vary depending on the size of the community in question, and the ways in which trustworthiness can be established and trusting warranted vary with these mechanisms. Mechanisms that work in dyads and small communities are often unavailable in the context of trusting an institution or branch of government. Establishing trust on this larger social scale therefore requires new or modified mechanisms. In the third section of the paper, we recommend three policies that – we argue – tend to make institutions more trustworthy and to reliably signal that trustworthiness to the public. First, they should ensure that their decision-making processes are as open and transparent as possible. Second, they should make efforts to engage stakeholders in dialogue with decision-makers such as managers, members of the C-Suite, and highly-placed policy-makers. Third, they should foster diversity – gender, ethnicity, age, socioeconomic background, disability, etc. – in their workforce at all levels, but especially in management and positions of power. We conclude by discussing the warrant for distrust in institutions that do not adopt these policies, which we contend is especially pertinent for people who belong to groups that have historically faced (and in many cases still do face) oppression. (shrink)

Previous simulation models have found positive effects of cognitive diversity on group performance, but have not explored effects of diversity in demographics (e.g., gender, ethnicity). In this paper, we present an agent-based model that captures two empirically supported hypotheses about how demographic diversity can improve group performance. The results of our simulations suggest that, even when social identities are not associated with distinctive task-related cognitive resources, demographic diversity can, in certain circumstances, benefit collective performance by counteracting two types of conformity (...) that can arise in homogeneous groups: those relating to group-based trust and those connected to normative expectations towards in-groups. (shrink)

In this paper, I articulate and defend a conception of trust that solves what I call “the trickster problem.” The problem results from the fact that many accounts of trust treat it similar to, or identical with, relying on someone’s good will. But a trickster could rely on your good will to get you to go along with his scheme, without trusting you to do so. Recent philosophical accounts of trust aim to characterize what it is for (...) one person to trust another so as to avoid this problem, but no extant account successfully does so. I argue that connecting trust to important, normatively defined relationships like friendship, romantic partnerships and parenting shows us something important about trust. The clearest cases of trust are found within the confines of normatively defined relationships like these, suggesting that there is a normative element to trust. Trusting someone involves not just believing that another person’s good will covers your interactions. Trusting involves believing that, at least in a certain domain of interaction, you are entitled to rely on that person’s good will. This account solves the trickster problem, because a trickster is not entitled to his victim’s good will. (shrink)

Our life with art is suffused with trust. We don’t just trust one another’s aesthetic testimony; we trust one another’s aesthetic actions. Audiences trust artists to have made it worth their while; artists trust audiences to put in the effort. Without trust, audiences would have little reason to put in the effort to understand difficult and unfamiliar art. I offer a theory of aesthetic trust, which highlights the importance of trust in aesthetic (...) sincerity. We trust in another’s aesthetic sincerity when we rely on them to fulfill their commitments to act for aesthetic reasons — rather than for, say, financial, social, or political reasons. We feel most thoroughly betrayed by an artist, not when they make bad art, but when they sell out. This teaches us something about the nature of trust in general. According to many standard theories, trust involves thinking the trusted to be cooperative or good-natured. But trust in aesthetic sincerity is different. We trust artists to be true to their own aesthetic sensibility, which might involve selfishly ignoring their audience’s needs. Why do we care so much about an artist’s sincerity, rather than merely trusting them to make good art? We emphasize sincerity when wish to encourage originality, rather than to demand success along predictable lines. And we ask for sincerity when our goal is to discover a shared sensibility. In moral life, we often try to force convergence through coordinated effort. But in aesthetic life, we often hope for the lovely discovery that our sensibilities were similar all along. And for that we need to ask for sincerity, rather than overt coordination. (shrink)

Technology is a practically indispensible means for satisfying one’s basic interests in all central areas of human life including nutrition, habitation, health care, entertainment, transportation, and social interaction. It is impossible for any one person, even a well-trained scientist or engineer, to know enough about how technology works in these different areas to make a calculated choice about whether to rely on the vast majority of the technologies she/he in fact relies upon. Yet, there are substantial risks, uncertainties, and unforeseen (...) practical consequences associated with the use of technological artifacts and systems. The salience of technological failure (both catastrophic and mundane), as well as technology’s sometimes unforeseeable influence on our behavior, makes it relevant to wonder whether we are really justified as individuals in our practical reliance on technology. Of course, even if we are not justified, we might nonetheless continue in our technological reliance, since the alternatives might not be attractive or feasible. In this chapter I argue that a conception of trust in technological artifacts and systems is plausible and helps us understand what is at stake philosophically in our reliance on technology. Such an account also helps us understand the relationship between trust and technological risk and the ethical obligations of those who design, manufacture, and deploy technological artifacts. (shrink)

A common criticism of the use of algorithms in criminal justice is that algorithms and their determinations are in some sense ‘opaque’—that is, difficult or impossible to understand, whether because of their complexity or because of intellectual property protections. Scholars have noted some key problems with opacity, including that opacity can mask unfair treatment and threaten public accountability. In this paper, we explore a different but related concern with algorithmic opacity, which centers on the role of public trust in (...) grounding the legitimacy of criminal justice institutions. We argue that algorithmic opacity threatens the trustworthiness of criminal justice institutions, which in turn threatens their legitimacy. We first offer an account of institutional trustworthiness before showing how opacity threatens to undermine an institution’s trustworthiness. We then explore how threats to trustworthiness affect institutional legitimacy. Finally, we offer some policy recommendations to mitigate the threat to trustworthiness posed by the opacity problem. (shrink)

This paper focuses on the epistemic situation one faces when using a Large Language Model based chatbot like ChatGPT: When reading the output of the chatbot, how should one decide whether or not to believe it? By surveying strategies we use with other, more familiar sources of information, I argue that chatbots present a novel challenge. This makes the question of how one could trust a chatbot especially vexing.

In this paper we develop a systematic account of trust in mathematics based on Katherine Hawley’s commitment account of trust. We focus on two related clusters of trust relationships in mathematical practice: (i) trusting relationships that are established when someone claims to have proven a result, and (ii) trusting others to check specific putative proofs. The image of mathematics that will emerge from our analysis is one of a deeply social discipline, where trust and its absence (...) are an important component of how justification goes from individual mathematicians to the mathematical community, and the social mechanisms of checking proofs. (shrink)