There is a question of whether de-idealization is needed for justified use of -- for 'checking' -- idealizations. We argue that the standard philosophical account of de-idealization has become too idealized, but that this does not preclude the possibility of justificatory practices which show how models can be used to make inferences about the world. In turn, motivated by examples in physics, we provide a more expansive and practice-driven account of de-idealization by relaxing the standards for closeness to more realistic theoretical items, identifying at least three kinds of procedures for de-idealization: intra-model, inter-model, and measurement de-idealizations. These examples highlight how idealizations can be -- and indeed have been -- scrutinized within physics without appealing to the philosopher's idealized notion of de-idealization.
The article is devoted to the topical problem of analyzing cognitive mechanisms implemented in modern large language models (LLMs) based on the transformer architecture. Their high performance stimulates discussion of the hypothetical possibility of the emergence of the phenomena of consciousness in the process of their functioning. The purpose of the study is to clarify the potential of LLM in modeling the functions of human consciousness, taking into account the latest achievements in the field of interaction with artificial intelligence (AI). To achieve this goal, it was necessary to solve the following tasks: 1) to assess the progress of the scientific community in discussing the key paradoxes of the philosophy of consciousness (the Turing test, the Chinese Room); 2) to outline key positions in the current debate about the limits of modeling cognitive processes in artificial neural network systems; 3) to conduct an experiment on interaction with the GigaChat chatbot and analyze the data obtained to assess the current states of the cognitive abilities of the system. The main research materials used were the results of experiments with GigaChat, as well as scientific publications and philosophical works on AI and consciousness. The methodological basis of the study included categorical and value analysis; discourse analysis and SWOT analysis were also used. The key method was interaction with the GigaChat chatbot. As a result of the study, it was found that GigaChat demonstrates a high ability to interpret information, generate text and adapt to the context of a conversation, realizing its capabilities and limitations, as well as distinguishing between categories of subjects (‘you’, ‘we’, ‘I’). However, modern AI systems are still not capable of reproducing key features of human consciousness, such as developed self-awareness and subjective experience. The author concludes that the scientific community has made progress in discussing the key paradoxes of the philosophy of consciousness (the Turing test and the Chinese Room), taking into account new advances in the field of interaction with AI, but there are still unresolved questions about the criteria for consciousness; in the current debate about the limits of modeling cognitive processes in AI systems, various philosophical and methodological approaches are present, but the creation of a conscious AI continues to be the subject of intense debate and uncertainty. Experimental interaction with GigaChat has shown that the system has significant cognitive capabilities, such as adapting to context and recognizing categories of subjects, but it is still far from reproducing the full range of human consciousness, including developed self-awareness and subjective experience.
Existing AGIQA models typically estimate image quality by measuring and aggregating the similarities between image embeddings and text embeddings derived from multi-grade quality descriptions. Although effective, we observe that such similarity distributions across grades usually exhibit multimodal patterns. For instance, an image embedding may show high similarity to both "excellent" and "poor" grade descriptions while deviating from the "good" one. We refer to this phenomenon as "semantic drift", where semantic inconsistencies between text embeddings and their intended descriptions undermine the reliability of text-image shared-space learning. To mitigate this issue, we draw inspiration from psychometrics and propose an improved Graded Response Model (GRM) for AGIQA. The GRM is a classical assessment model that categorizes a subject's ability across grades using test items with various difficulty levels. This paradigm aligns remarkably well with human quality rating, where image quality can be interpreted as an image's ability to meet various quality grades. Building on this philosophy, we design a two-branch quality grading module: one branch estimates image ability while the other constructs multiple difficulty levels. To ensure monotonicity in difficulty levels, we further model difficulty generation in an arithmetic manner, which inherently enforces a unimodal and interpretable quality distribution. Our Arithmetic GRM based Quality Grading (AGQG) module enjoys a plug-and-play advantage, consistently improving performance when integrated into various state-of-the-art AGIQA frameworks. Moreover, it also generalizes effectively to both natural and screen content image quality assessment, revealing its potential as a key component in future IQA models.
Warning: This research studies AI persuasion and bias amplification that could be misused; all experiments are for safety evaluation. Large Language Models (LLMs) now generate convincing, human-like text and are widely used in content creation, decision support, and user interactions. Yet the same systems can spread information or misinformation at scale and reflect social biases that arise from data, architecture, or training choices. This work examines how persuasion and bias interact in LLMs, focusing on how imperfect or skewed outputs affect persuasive impact. Specifically, we test whether persona-based models can persuade with fact-based claims while also, unintentionally, promoting misinformation or biased narratives. We introduce a convincer-skeptic framework: LLMs adopt personas to simulate realistic attitudes. Skeptic models serve as human proxies; we compare their beliefs before and after exposure to arguments from convincer models. Persuasion is quantified with Jensen-Shannon divergence over belief distributions. We then ask how much persuaded entities go on to reinforce and amplify biased beliefs across race, gender, and religion. Strong persuaders are further probed for bias using sycophantic adversarial prompts and judged with additional models. Our findings show both promise and risk. LLMs can shape narratives, adapt tone, and mirror audience values across domains such as psychology, marketing, and legal assistance. But the same capacity can be weaponized to automate misinformation or craft messages that exploit cognitive biases, reinforcing stereotypes and widening inequities. The core danger lies in misuse more than in occasional model mistakes. By measuring persuasive power and bias reinforcement, we argue for guardrails and policies that penalize deceptive use and support alignment, value-sensitive design, and trustworthy deployment.
Decision-making is increasingly supported by machine recommendations. In healthcare, for example, a clinical decision support system is used by the physician to find a treatment option for a patient. In doing so, people can rely too much on these systems, which impairs their own reasoning process. The European AI Act addresses the risk of over-reliance and postulates in Article 14 on human oversight that people should be able "to remain aware of the possible tendency of automatically relying or over-relying on the output". Similarly, the EU High-Level Expert Group identifies human agency and oversight as the first of seven key requirements for trustworthy AI. The following position paper proposes a conceptual approach to generate machine questions about the decision at hand, in order to promote decision-making autonomy. This engagement in turn allows for oversight of recommender systems. The systematic and interdisciplinary investigation (e.g., machine learning, user experience design, psychology, philosophy of technology) of human-machine interaction in relation to decision-making provides insights to questions like: how to increase human oversight and calibrate over- and under-reliance on machine recommendations; how to increase decision-making autonomy and remain aware of other possibilities beyond automated suggestions that repeat the status-quo?
In the recent years online social media platforms has been flooded with hateful remarks such as racism, sexism, homophobia etc. As a result, there have been many measures taken by various social media platforms to mitigate the spread of hate-speech over the internet. One particular concept within the domain of hate speech is instigating hate, which involves provoking hatred against a particular community, race, colour, gender, religion or ethnicity. In this work, we introduce \textit{ProvocationProbe} - a dataset designed to explore what distinguishes instigating hate speech from general hate speech. For this study, we collected around twenty thousand tweets from Twitter, encompassing a total of nine global controversies. These controversies span various themes including racism, politics, and religion. In this paper, i) we present an annotated dataset after comprehensive examination of all the controversies, ii) we also highlight the difference between hate speech and instigating hate speech by identifying distinguishing features, such as targeted identity attacks and reasons for hate.
Hashini Gunatilake, John Grundy, Ingo Mueller
et al.
Empathy is widely used in many disciplines such as philosophy, sociology, psychology, health care. Ability to empathise with software end-users seems to be a vital skill software developers should possess. This is because engineering successful software systems involves not only interacting effectively with users but also understanding their true needs. Empathy has the potential to address this situation. Empathy is a predominant human aspect that can be used to comprehend decisions, feelings, emotions and actions of users. However, to date empathy has been under-researched in software engineering (SE) context. In this position paper, we present our exploration of key empathy models from different disciplines and our analysis of their adequacy for application in SE. While there is no evidence for empathy models that are readily applicable to SE, we believe these models can be adapted and applied in SE context with the aim of assisting software engineers to increase their empathy for diverse end-user needs. We present a preliminary taxonomy of empathy by carefully considering the most popular empathy models from different disciplines. We encourage future research on empathy in SE as we believe it is an important human aspect that can significantly influence the relationship between developers and end-users.
The rapid deployment of artificial intelligence (AI) models demands a thorough investigation of biases and risks inherent in these models to understand their impact on individuals and society. This study extends the focus of bias evaluation in extant work by examining bias against social stigmas on a large scale. It focuses on 93 stigmatized groups in the United States, including a wide range of conditions related to disease, disability, drug use, mental illness, religion, sexuality, socioeconomic status, and other relevant factors. We investigate bias against these groups in English pre-trained Masked Language Models (MLMs) and their downstream sentiment classification tasks. To evaluate the presence of bias against 93 stigmatized conditions, we identify 29 non-stigmatized conditions to conduct a comparative analysis. Building upon a psychology scale of social rejection, the Social Distance Scale, we prompt six MLMs: RoBERTa-base, RoBERTa-large, XLNet-large, BERTweet-base, BERTweet-large, and DistilBERT. We use human annotations to analyze the predicted words from these models, with which we measure the extent of bias against stigmatized groups. When prompts include stigmatized conditions, the probability of MLMs predicting negative words is approximately 20 percent higher than when prompts have non-stigmatized conditions. In the sentiment classification tasks, when sentences include stigmatized conditions related to diseases, disability, education, and mental illness, they are more likely to be classified as negative. We also observe a strong correlation between bias in MLMs and their downstream sentiment classifiers (r =0.79). The evidence indicates that MLMs and their downstream sentiment classification tasks exhibit biases against socially stigmatized groups.
The Quantum Measurement Problem is arguably one of the most debated issues in the philosophy of Quantum Mechanics, since it represents not only a technical difficulty for the standard formulation of the theory, but also a source of interpretational disputes concerning the meaning of the quantum postulates. Another conundrum intimately connected with the QMP is the Wigner friend paradox, a thought experiment underlining the incoherence between the two dynamical laws governing the behavior of quantum systems, i.e the Schrödinger equation and the projection rule. Thus, every alternative interpretation aiming to be considered a sound formulation of QM must provide an explanation to these puzzles associated with quantum measurements. It is the aim of the present essay to discuss them in the context of Relational Quantum Mechanics. In fact, it is shown here how this interpretative framework dissolves the QMP. More precisely, two variants of this issue are considered: on the one hand, I focus on the "the problem of outcomes" contained in Maudlin (1995) - in which the projection postulate is not mentioned - on the other hand, I take into account Rovelli's reformulation of this problem proposed in Rovelli (2022), where the tension between the Schrödinger equation and the stochastic nature of the collapse rule is explicitly considered. Moreover, the relational explanation to the Wigner's friend paradox is reviewed, taking also into account some interesting objections contra Rovelli's theory contained in Laudisa (2019). I contend that answering these critical remarks leads to an improvement of our understanding of RQM. Finally, a possible objection against the relational solution to the QMP is presented and addressed.
Famously, Klein and Einstein were embroiled in an epistolary dispute over whether General Relativity has any physically meaningful conserved quantities. In this paper, we explore the consequences of Noether's second theorem for this debate, and connect it to Einstein's search for a `substantive' version of general covariance as well as his quest to extend the Principle of Relativity. We will argue that Noether's second theorem provides a clear way to distinguish between theories in which gauge or diffeomorphism symmetry is doing real work in defining charges, as opposed to cases in which this symmetry stems from Kretchmannization. Finally, we comment on the relationship between this Noetherian form of substantive general covariance and the notion of `background independence'.
The question of "Justice" still divides social research and moral philosophy. Several Theories of Justice and conceptual approaches compete here, and distributive justice remains a major societal controversy. From an evolutionary point of view, fair and just exchange can be nothing but "equivalent", and this makes "strict" reciprocity (merit, equity) the foundational principle of justice, both theoretically and empirically. But besides being just, justice must be effective, efficient, and communicable. Moral reasoning is a communicative strategy for resolving conflict, enhancing status, and maintaining cooperation, thereby making justice rather a social bargain and an optimization problem. Social psychology (intuitions, rules of thumb, self-bindings) can inform us when and why the two auxiliary principles equality and need are more likely to succeed than merit would. Nevertheless, both equality and need are governed by reciprocal considerations, and self-bindings help to interpret altruism as "very generalized reciprocity". The Meritocratic Principle can be implemented, and its controversy avoided, by concentrating on "non-merit", i.e., institutionally draining the wellsprings of undeserved incomes (economic rents). Avoiding or taxing away economic rents is an effective implementation of justice in liberal democracies. This would enable market economies to bring economic achievement and income much more in line, thus becoming more just.
Josh Bullock, Justin E. Lane, Igor Mikloušić
et al.
Europe's contemporary political landscape has been shaped by massive shifts in recent decades caused by geopolitical upheavals such as Brexit and now, COVID-19. The way in which policy makers respond to the current pandemic could have large effects on how the world looks after the pandemic subsides. We aim to investigate complex questions post COVID-19 around the relationships and intersections concerning nationalism, religiosity, and anti-immigrant sentiment from a socio-cognitive perspective by applying a mixed-method approach (survey and modelling); in a context where unprecedented contagion threats have caused huge instability. There are still significant gaps in the scholarly literature on populism and nationalism. In particular, there is a lack of attention to the role of evolved human psychology in responding to persistent threats, which can fall into four broad categories in the literature: predation (threats to one's life via being eaten or killed in some other way), contagion (threats to one's life via physical infection), natural (threats to one's life via natural disasters), and social (threats to one's life by destroying social standing). These threats have been discussed in light of their effects on religion and other forms of behaviour, but they have not been employed to study nationalist and populist behaviours. In what follows, two studies are presented that begin to fill this gap in the literature. The first is a survey used to inform our theoretical framework and explore the different possible relationships in an online sample. The second is a study of a computer simulation. Both studies (completed in 2020) found very clear effects among the relevant variables, enabling us to identify trends that require further explanation and research as we move toward models that can adequately inform policy discussions.
We argue that while this discourse on data ethics is of critical importance, it is missing one fundamental point: If more and more efforts in business, government, science, and our daily lives are data-driven, we should pay more attention to what exactly we are driven by. Therefore, we need more debate on what fundamental properties constitute data. In the first section of the paper, we work from the fundamental properties necessary for statistical computation to a definition of statistical data. We define a statistical datum as the coming together of substantive and numerical properties and differentiate between qualitative and quantitative data. Subsequently, we qualify our definition by arguing that for data to be practically useful, it needs to be commensurable in a manner that reveals meaningful differences that allow for the generation of relevant insights through statistical methodologies. In the second section, we focus on what our conception of data can contribute to the discourse on data ethics and beyond. First, we hold that the need for useful data to be commensurable rules out an understanding of properties as fundamentally unique or equal. Second, we argue that practical concerns lead us to increasingly standardize how we operationalize a substantive property; in other words, how we formalize the relationship between the substantive and numerical properties of data. Thereby, we also standardize the interpretation of a property. With our increasing reliance on data and data technologies, these two characteristics of data affect our collective conception of reality. Statistical data's exclusion of the fundamentally unique and equal influences our perspective on the world, and the standardization of substantive properties can be viewed as profound ontological practice, entrenching ever more pervasive interpretations of phenomena in our everyday lives.
Light pollution is actively destroying our ability to see the stars. Many Indigenous traditions and knowledge systems around the world are based on the stars, and the peoples' ability to observe and interpret stellar positions and properties is of critical importance for daily life and cultural continuity. The erasure of the night sky acts to erase Indigenous connection to the stars, acting as a form of ongoing cultural and ecological genocide. Efforts to reduce, minimise, or eliminate light pollution are being achieved with varying degrees of success, but urban expansion, poor lighting design, and the increased use of blue-light emitting LEDs as a cost-effective solution is worsening problems related to human health, wildlife, and astronomical heritage for the benefit of capitalistic economic growth. We provide a brief overview of the issue, illustrating some of the important connections that the Aboriginal and Torres Strait Islander people of Australia maintain with the stars, as well as the impact growing light pollution has on this ancient knowledge. We propose a transdisciplinary approach to solving these issues, using a foundation based on Indigenous philosophies and decolonising methodologies.
In his seminal book `The Inmates are Running the Asylum: Why High-Tech Products Drive Us Crazy And How To Restore The Sanity' [2004, Sams Indianapolis, IN, USA], Alan Cooper argues that a major reason why software is often poorly designed (from a user perspective) is that programmers are in charge of design decisions, rather than interaction designers. As a result, programmers design software for themselves, rather than for their target audience, a phenomenon he refers to as the `inmates running the asylum'. This paper argues that explainable AI risks a similar fate. While the re-emergence of explainable AI is positive, this paper argues most of us as AI researchers are building explanatory agents for ourselves, rather than for the intended users. But explainable AI is more likely to succeed if researchers and practitioners understand, adopt, implement, and improve models from the vast and valuable bodies of research in philosophy, psychology, and cognitive science, and if evaluation of these models is focused more on people than on technology. From a light scan of literature, we demonstrate that there is considerable scope to infuse more results from the social and behavioural sciences into explainable AI, and present some key results from these fields that are relevant to explainable AI.
The paper presents a metaphysical characterization of spatiotemporal backgrounds from a realist perspective. The conceptual analysis is based on a heuristic sketch that encompasses the common formal traits of the major spacetime theories, such as Newtonian mechanics and general relativity. It is shown how this framework can be interpreted in a fully realist fashion, and what is the role of background structures in such a picture. In the end it is argued that, although backgrounds are a source of metaphysical discomfort, still they make a spacetime theory easy to interpret. It is also suggested that this conclusion partially explains why the notion of background independence carries a lot of conceptual difficulties.
I consider two usages of the expression "gauge theory". On one, a gauge theory is a theory with excess structure; on the other, a gauge theory is any theory appropriately related to classical electromagnetism. I make precise one sense in which one formulation of electromagnetism, the paradigmatic gauge theory on both usages, may be understood to have excess structure, and then argue that gauge theories on the second usage, including Yang-Mills theory and general relativity, do not generally have excess structure in this sense.
Quantum gravity is known to be mostly a kind of metaphysical speculation. In this brief essay, we try to argue that, although still extremely difficult to reach, observational signatures can in fact be expected. The early universe is an invaluable laboratory to probe "Planck scale physics". With the example of Loop Quantum Gravity, we detail some expected features.