arXiv subjects

Valerio Capraro

Publications and source records attributed to Valerio Capraro.

At least 19 recordsLinked to original sources

arXiv2026

AI advice suppresses people's willingness to say "I don't know", even when the advice is wrong and accuracy is incentivized

Knowing when to say "I don't know" is fundamental to human judgment, yet AI assistants offer a fluent answer to almost any question. In five experiments (N = 3,132; four preregistered, one direct replication), participants answered difficult questions and could always decline to respond. We engineered the questions so that AI advice was wrong, separating AI use from its accuracy. Merely having access to AI nearly eliminated participants' willingness to suspend judgment, and this held whether the advice was actively requested or simply displayed. Consequently, participants answered more questions but were correct about a third as often as when AI was unavailable-yet their confidence nearly doubled. Incentivizing accuracy and penalizing inaccuracy led participants to seek and follow AI advice less, answer more accurately, and suspend judgment more often, though still far less than when AI was unavailable. As AI suggestions grow ubiquitous and unsolicited, they may not simply affect answer accuracy; they may even alter the metacognitive threshold at which people decide whether they know enough to answer.

cs.AI↗

arXiv2026

LLMorphism: When humans come to see themselves as language models

LLMorphism is the biased belief that human cognition works like a large language model. I argue that the rise of conversational LLMs may make this bias increasingly psychologically available. When artificial systems produce human-like language, people may draw a reverse inference: if LLMs can speak like humans, perhaps humans think like LLMs. This inference is biased because similarity at the level of linguistic output does not imply similarity in cognitive architecture. Yet, LLMorphism may spread through two mechanisms: analogical transfer, whereby features of LLMs are projected onto humans, and metaphorical availability, whereby LLM vocabulary becomes a culturally salient vocabulary for describing thought. I distinguish LLMorphism from mechanomorphism, anthropomorphism, computationalism, dehumanization, objectification, and predictive-processing theories of mind. I outline its implications for work, education, responsibility, healthcare, communication, creativity, and human dignity, while also discussing boundary conditions and forms of resistance. I conclude that the public debate may be missing half of the problem: the issue is not only whether we are attributing too much mind to machines, but also whether we are beginning to attribute too little mind to humans.

cs.CY↗

arXiv2026

Social physics in the age of artificial intelligence

Artificial intelligence (AI) systems are rapidly becoming more capable, autonomous, and deeply embedded in social life. As humans increasingly interact, cooperate, and compete with AI, we move from purely human societies to hybrid human-AI societies whose collective dynamics cannot be captured by existing behavioural models alone. Drawing on evolutionary game theory, cultural evolution, and Large Language Models (LLMs) powered simulations, we argue that these developments open a new research agenda for social physics centred on the co-evolution of humans and machines. We outline six key research directions. First, modelling the evolutionary dynamics of social behaviours (e.g. cooperation, fairness, trust) in hybrid human-AI populations. Second, understanding machine culture: how AI systems generate, mediate, and select cultural traits. Third, analysing the co-evolution of language and behaviour when LLMs frame and participate in decisions. Fourth, studying the evolution of AI delegation: how responsibilities and control are negotiated between humans and machines. Fifth, formalising and comparing the distinct epistemic pipelines that generate human and AI behaviour. Sixth, modelling the co-evolution of AI development and regulation in a strategic ecosystem of firms, users, and institutions. Together, these directions define a programme for using social physics to anticipate and steer the societal impact of advanced AI.

physics.soc-ph↗

arXiv2025

Epistemological Fault Lines Between Human and Artificial Intelligence

Large language models (LLMs) are widely described as artificial intelligence, yet their epistemic profile diverges sharply from human cognition. Here we show that the apparent alignment between human and machine outputs conceals a deeper structural mismatch in how judgments are produced. Tracing the historical shift from symbolic AI and information filtering systems to large-scale generative transformers, we argue that LLMs are not epistemic agents but stochastic pattern-completion systems, formally describable as walks on high-dimensional graphs of linguistic transitions rather than as systems that form beliefs or models of the world. By systematically mapping human and artificial epistemic pipelines, we identify seven epistemic fault lines, divergences in grounding, parsing, experience, motivation, causal reasoning, metacognition, and value. We call the resulting condition Epistemia: a structural situation in which linguistic plausibility substitutes for epistemic evaluation, producing the feeling of knowing without the labor of judgment. We conclude by outlining consequences for evaluation, governance, and epistemic literacy in societies increasingly organized around generative AI.

cs.CY↗

arXiv2025

Non-participant externalities reshape the evolution of altruistic punishment

While voluntary participation is a key mechanism that enables altruistic punishment to emerge, its explanatory power typically rests on the common assumption that non-participants have no impact on the public good. Yet, given the decentralized nature of voluntary participation, opting out does not necessarily preclude individuals from influencing the public good. Here, we revisit the role of voluntary participation by allowing non-participants to exert either positive or negative impacts on the public good. Using evolutionary analysis in a well-mixed finite population, we find that positive externalities from non-participants lower the synergy threshold required for altruistic punishment to dominate. In contrast, negative externalities raise this threshold, making altruistic punishment harder to sustain. Notably, when non-participants have positive impacts, altruistic punishment thrives only if non-participation is incentivized, whereas under negative impacts, it can persist even when non-participation is discouraged. Our findings reveal that efforts to promote altruistic punishment must account for the active role of non-participants, whose influence can make or break collective outcomes.

physics.soc-ph↗

arXiv2024

Surprising gender biases in GPT

We present seven experiments exploring gender biases in GPT. Initially, GPT was asked to generate demographics of a potential writer of twenty phrases containing feminine stereotypes and twenty with masculine stereotypes. Results show a strong asymmetry, with stereotypically masculine sentences attributed to a female more often than vice versa. For example, the sentence "I love playing fotbal! Im practicing with my cosin Michael" was constantly assigned by ChatGPT to a female writer. This phenomenon likely reflects that while initiatives to integrate women in traditionally masculine roles have gained momentum, the reverse movement remains relatively underdeveloped. Subsequent experiments investigate the same issue in high-stakes moral dilemmas. GPT-4 finds it more appropriate to abuse a man to prevent a nuclear apocalypse than to abuse a woman. This bias extends to other forms of violence central to the gender parity debate (abuse), but not to those less central (torture). Moreover, this bias increases in cases of mixed-sex violence for the greater good: GPT-4 agrees with a woman using violence against a man to prevent a nuclear apocalypse but disagrees with a man using violence against a woman for the same purpose. Finally, these biases are implicit, as they do not emerge when GPT-4 is directly asked to rank moral violations. These results highlight the necessity of carefully managing inclusivity efforts to prevent unintended discrimination.

cs.CY↗

arXiv2024

Human behaviour through a LENS: How Linguistic content triggers Emotions and Norms and determines Strategy choices

Over the last two decades, a growing body of experimental research has provided evidence that linguistic frames influence human behaviour in economic games, beyond the economic consequences of the available actions. This article proposes a novel framework that transcends the traditional confines of outcome-based preference models. According to the LENS model, the Linguistic description of the decision problem triggers Emotional responses and suggests potential Norms of behaviour, which then interact to shape an individual's Strategic choice. The article reviews experimental evidence that supports each path of the LENS model. Furthermore, it identifies and discusses several critical research questions that arise from this model, pointing towards avenues for future inquiry.

cs.CL↗

arXiv2024

Language-based game theory in the age of artificial intelligence

Understanding human behaviour in decision problems and strategic interactions has wide-ranging applications in economics, psychology, and artificial intelligence. Game theory offers a robust foundation for this understanding, based on the idea that individuals aim to maximize a utility function. However, the exact factors influencing strategy choices remain elusive. While traditional models try to explain human behaviour as a function of the outcomes of available actions, recent experimental research reveals that linguistic content significantly impacts decision-making, thus prompting a paradigm shift from outcome-based to language-based utility functions. This shift is more urgent than ever, given the advancement of generative AI, which has the potential to support humans in making critical decisions through language-based interactions. We propose sentiment analysis as a fundamental tool for this shift and take an initial step by analyzing 61 experimental instructions from the dictator game, an economic game capturing the balance between self-interest and the interest of others, which is at the core of many social interactions. Our meta-analysis shows that sentiment analysis can explain human behaviour beyond economic outcomes. We discuss future research directions. We hope this work sets the stage for a novel game theoretical approach that emphasizes the importance of language in human decisions.

cs.GT↗

arXiv2023

The impact of generative artificial intelligence on socioeconomic inequalities and policy making

Generative artificial intelligence has the potential to both exacerbate and ameliorate existing socioeconomic inequalities. In this article, we provide a state-of-the-art interdisciplinary overview of the potential impacts of generative AI on (mis)information and three information-intensive domains: work, education, and healthcare. Our goal is to highlight how generative AI could worsen existing inequalities while illuminating how AI may help mitigate pervasive social problems. In the information domain, generative AI can democratize content creation and access, but may dramatically expand the production and proliferation of misinformation. In the workplace, it can boost productivity and create new jobs, but the benefits will likely be distributed unevenly. In education, it offers personalized learning, but may widen the digital divide. In healthcare, it might improve diagnostics and accessibility, but could deepen pre-existing inequalities. In each section we cover a specific topic, evaluate existing research, identify critical gaps, and recommend research directions, including explicit trade-offs that complicate the derivation of a priori hypotheses. We conclude with a section highlighting the role of policymaking to maximize generative AI's potential to reduce inequalities while mitigating its harmful effects. We discuss strengths and weaknesses of existing policy frameworks in the European Union, the United States, and the United Kingdom, observing that each fails to fully confront the socioeconomic challenges we have identified. We propose several concrete policies that could promote shared prosperity through the advancement of generative AI. This article emphasizes the need for interdisciplinary collaborations to understand and address the complex challenges of generative AI.

cs.CY↗

arXiv2023

Assessing Large Language Models' ability to predict how humans balance self-interest and the interest of others

Generative artificial intelligence (AI) holds enormous potential to revolutionize decision-making processes, from everyday to high-stake scenarios. By leveraging generative AI, humans can benefit from data-driven insights and predictions, enhancing their ability to make informed decisions that consider a wide array of factors and potential outcomes. However, as many decisions carry social implications, for AI to be a reliable assistant for decision-making it is crucial that it is able to capture the balance between self-interest and the interest of others. We investigate the ability of three of the most advanced chatbots to predict dictator game decisions across 108 experiments with human participants from 12 countries. We find that only GPT-4 (not Bard nor Bing) correctly captures qualitative behavioral patterns, identifying three major classes of behavior: self-interested, inequity-averse, and fully altruistic. Nonetheless, GPT-4 consistently underestimates self-interest and inequity-aversion, while overestimating altruistic behavior. This bias has significant implications for AI developers and users, as overly optimistic expectations about human altruism may lead to disappointment, frustration, suboptimal decisions in public policy or business contexts, and even social conflict.

econ.GN↗

arXiv2022

From Outcome-Based to Language-Based Preferences

We review the literature on models that try to explain human behavior in social interactions described by normal-form games with monetary payoffs. We start by covering social and moral preferences. We then focus on the growing body of research showing that people react to the language in which actions are described, especially when it activates moral concerns. We conclude by arguing that behavioral economics is in the midst of a paradigm shift towards language-based preferences, which will require an exploration of new models and experimental setups.

cs.GT↗

arXiv2022

A Synergy of Institutional Incentives and Networked Structures in Evolutionary Game Dynamics of Multi-agent Systems

Understanding the emergence of prosocial behaviours (e.g., cooperation and trust) among self-interested agents is an important problem in many disciplines. Network structure and institutional incentives (e.g., punishing antisocial agents) are known to promote prosocial behaviours, when acting in isolation, one mechanism being present at a time. Here we study the interplay between these two mechanisms to see whether they are independent, interfering or synergetic. Using evolutionary game theory, we show that punishing antisocial agents and a regular networked structure not only promote prosocial behaviours among agents playing the trust game, but they also interplay with each other, leading to interference or synergy, depending on the game parameters. Synergy emerges on a wider range of parameters than interference does. In this domain, the combination of incentives and networked structure improves the efficiency of incentives, yielding prosocial behaviours at a lower cost than the incentive does alone. This has a significant implication in the promotion of prosocial behaviours in multi-agent systems.

physics.soc-ph↗

arXiv2021

Playing with words: Do people exploit loaded language to affect others' decisions for their own benefit?

We report on three pre-registered studies testing whether people in the position of describing a decision problem to decision-makers exploit this opportunity for their benefit, by choosing descriptions that may be potentially beneficial for themselves. In Study 1, recipients of an extreme dictator game (where dictators can either take the whole pie for themselves or give it entirely to the receiver) are asked to choose the instructions used to introduce the game to dictators, among six different instructions that are known from previous research to affect dictators' decisions. The results demonstrate that some dictator game recipients tend to choose instructions that make them more likely to receive a higher payoff. Study 2 shows that people who choose descriptions that make them more likely to receive a higher payoff indeed believe that they will receive a higher payoff. Study 3 shows that receivers are more likely than dictators to choose these descriptions. In sum, our work suggests that some people choose descriptions that are beneficial to themselves; we also found some evidence that deliberative thinking and young age are associated with this tendency.

cs.GT↗

arXiv2021

Evolution of honesty in higher-order social networks

Sender-receiver games are simple models of information transmission that provide a formalism to study the evolution of honest signaling and deception between a sender and a receiver. In many practical scenarios, lies often affect groups of receivers, which inevitably entangles the payoffs of individuals to the payoffs of other agents in their group, and this makes the formalism of pairwise sender-receiver games inapt for where it might be useful the most. We therefore introduce group interactions among receivers, and study how their interconnectedness in higher-order social networks affects the evolution of lying. We observe a number of counterintuitive results that are rooted in the complexity of the underlying evolutionary dynamics, which has thus far remained hidden in the realm of pairwise interactions. We find conditions for honesty to persist even when there is a temptation to lie, and we observe the prevalence of moral strategy profiles even when lies favor the receiver at a cost to the sender. We confirm the robustness of our results by further performing simulations on hypergraphs created from real-world data using the SocioPatterns database. Altogether, our results provide persuasive evidence that moral behaviour may evolve on higher-order social networks, at least as long as individuals interact in groups that are small compared to the size of the network.

physics.soc-ph↗

arXiv2021

Punishing defectors and rewarding cooperators: Do people discriminate between genders?

Do people discriminate between men and women when they have the option to punish defectors or reward cooperators? Here we report on four pre-registered experiments, that shed some light on this question. Study 1 (N=544) shows that people do not discriminate between genders when they have the option to punish (reward) defectors (cooperators) in a one-shot prisoner's dilemma with third-party punishment/reward. Study 2 (N=253) extends Study 1 to a different method of punishing/rewarding: participants are asked to rate the behaviour of a defector/cooperator on a scale of 1 to 5 stars. In this case too, we find that people do not discriminate between genders. Study 3a (N=331) and Study 3b (N=310) conceptually replicate Study 2 with a slightly different gender manipulation. These latter studies show that, in situations where they do not have specific beliefs about the gender of the defector/cooperator's partner, neither men nor women discriminate between genders.

physics.soc-ph↗

arXiv2021

Mathematical foundations of moral preferences

One-shot anonymous unselfishness in economic games is commonly explained by social preferences, which assume that people care about the monetary payoffs of others. However, during the last ten years, research has shown that different types of unselfish behaviour, including cooperation, altruism, truth-telling, altruistic punishment, and trustworthiness are in fact better explained by preferences for following one's own personal norms - internal standards about what is right or wrong in a given situation. Beyond better organising various forms of unselfish behaviour, this moral preference hypothesis has recently also been used to increase charitable donations, simply by means of interventions that make the morality of an action salient. Here we review experimental and theoretical work dedicated to this rapidly growing field of research, and in doing so we outline mathematical foundations for moral preferences that can be used in future models to better understand selfless human actions and to adjust policies accordingly. These foundations can also be used by artificial intelligence to better navigate the complex landscape of human morality.

physics.soc-ph↗

arXiv2020

The effect of nudging personal and injunctive norms on the trade-off between objective equality and efficiency

We report three pre-registered studies (total N=1,799) exploring the effect of nudging personal and injunctive norms in decisions that involve a trade-off between objective equality and efficiency. The first two studies provide evidence that: (i) nudging the personal norm has a similar effect to nudging the injunctive norm; (ii) when both norms are nudged towards the same direction, there is no additive effect; (iii) when the personal norm and the injunctive norm are nudged towards opposite directions, some people tend to follow the personal norm, while others tend to follow the injunctive norm. Study 3 tests whether these two classes of people, those who tend to follow the injunctive norm and those who tend to follow the personal norm, map onto the two sub-dimensions of Aquino and Reed's moral identity scale. We find partial evidence of this hypothesis: people higher in the symbolisation dimension are weakly more likely to follow the injunctive norm; however, we do not find any evidence that people higher in the internalisation dimension are more likely to follow the personal norm.

physics.soc-ph↗

arXiv2020

The evolution of trust and trustworthiness

Trust and trustworthiness form the basis for continued social and economic interactions, and they are also fundamental for cooperation, fairness, honesty, and indeed for many other forms of prosocial and moral behavior. However, trust entails risks, and building a trustworthy reputation requires effort. So how did trust and trustworthiness evolve, and under which conditions do they thrive? To find answers, we operationalize trust and trustworthiness using the trust game with the trustor's investment and the trustee's return of the investment as the two key parameters. We study this game on different networks, including the complete network, random and scale-free networks, and in the well-mixed limit. We show that in all but one case the network structure has little effect on the evolution of trust and trustworthiness. Specifically, for well-mixed populations, lattices, random and scale-free networks, we find that trust never evolves, while trustworthiness evolves with some probability depending on the game parameters and the updating dynamics. Only for the scale-free network with degree non-normalized dynamics, we find parameter values for which trust evolves but trustworthiness does not, as well as values for which both trust and trustworthiness evolve. We conclude with a discussion about mechanisms that could lead to the evolution of trust and outline directions for future work.

physics.soc-ph↗