arXiv subjects

Hirokazu Shirado

Publications and source records attributed to Hirokazu Shirado.

11 recordsLinked to original sources

arXiv2026

Modeling Duelling Contagions of True and False Information in the Face of Inherent Individual biases

Advanced digital communication has revolutionized how people create and consume information, making information diffusion an important topic of research for domains from public health to national security. Real-world scenarios of information diffusion often involve competing narratives - true and false - spreading simultaneously. We propose a novel agent-based co-diffusion model, grounded in "complex-contagion" and "spiral of silence" theories, to capture how network dynamics exploit cognitive biases to shape such interactions. Our findings reveal that manipulative narratives dominate when early spreaders hold them. These network dynamics further exploit inherent cognitive biases to amplify information diffusion regardless of veracity. Further, while favourable previous experience strengthen collective optimism, unfavourable experiences attenuate optimism only modestly. However, we found that early seeding of agents with lower self-censorship not only constrains the spread of manipulation but can also lead to dominance of well-informed populance. This has implications for policies that aim to facilitate healthier discourse, strengthen social cohesion, and ensure equitable access to reliable information.

cs.SI↗

arXiv2026

Faith in AI can narrow the futures individuals consider

Artificial intelligence (AI) predictions are increasingly used to inform human decisions. Here, using a behavioral implementation of the classic Newcomb's paradox in 1,305 participants, we show that AI predictions can also shape the reasoning people use to make a decision. In this paradigm, perceived predictive authority can alter how people reason about their future actions, leading them to forgo a guaranteed reward. Over 40% of participants treated AI as such a predictive authority about their own behavior, significantly increasing the odds of forgoing the guaranteed reward by a factor of 3.39 (95% CI: 2.45-4.70) and reducing earnings by 10.7-42.9%. The effect appeared across AI presentations and decision contexts and remained detectable even when predictions repeatedly failed. When people perceive AI as capable of predicting their personal behavior, the mere presence of AI predictions may shape their decision-making, narrowing the futures they consider.

cs.HC↗

arXiv2026

Systematic Failures in Collective Reasoning under Distributed Information in Multi-Agent LLMs

Multi-agent systems built on large language models (LLMs) are expected to enhance decision-making by pooling distributed information, yet systematically evaluating this capability has remained challenging. We introduce HiddenBench, a 65-task benchmark grounded in the Hidden Profile paradigm, which isolates collective reasoning under distributed information from individual reasoning ability. Evaluating 15 frontier LLMs, we find that multi-agent LLMs achieve only 30.1% accuracy under distributed information, compared to 80.7% accuracy for single agents given complete information. We trace this gap to a systematic failure mode: agents cannot recognize or act under latent information asymmetry -- they fail to reason about what others might know but have not yet expressed, leading to premature convergence on shared evidence while critical distributed facts remain unexplored. These failures persist across prompting strategies, communication depths, and group sizes -- and worsen as groups scale. While some models (e.g., Gemini-2.5-Flash/Pro) outperform others, neither model scale nor individual reasoning accuracy reliably predicts collective performance. We further show that this bottleneck is actionable: a lightweight structured communication protocol substantially improves collective reasoning across model families. Our results identify failures in collective information exploration in decision-making as a key limitation of multi-agent LLMs, and provide a theory-grounded, reproducible framework for diagnosing collective reasoning failures.

cs.CL↗

arXiv2026

WhatIf: Interactive Exploration of LLM-Powered Social Simulations for Policy Reasoning

Policymakers in domains such as emergency management, public health, and urban planning must make decisions under deep uncertainty, where outcomes depend on how large populations interpret information, coordinate, and adopt over time. Existing tools only partially support this process: tabletop exercises enable collaborative discussion but lack dynamic feedback, while computational simulations capture population dynamics but are designed for offline analysis. We present WhatIf, an interactive system that enables policymakers to steer, inspect, and compare LLM-powered social simulations in real time. Informed by a formative study in emergency preparedness planning, we derive four design requirements for interactive policy simulations: fluid steering, real-time scale, collaborative exploration, and multi-level interpretability. We developed WhatIf guided by these requirements and evaluated it with five preparedness professionals across three disaster evacuation scenarios. Our findings show that participants used the system as a space for iterative branching and comparison rather than evaluating fixed plans; reflected on tacit planning assumptions when agent behavior violated expectations; surfaced previously unrecognized planning vulnerabilities; and grounded their reasoning in inspectable agent-level cases rather than aggregate outputs alone. These findings suggest broader design implications for LLM-powered social simulation systems: designing such systems as interactive, shared reasoning environments -- rather than offline predictive tools -- can better support expert decision-making under deep uncertainty.

cs.HC↗

arXiv2026

What Makes LLM Agent Simulations Useful for Policy Practice? An Iterative Design Study in Emergency Preparedness

Policymakers must often act under conditions of deep uncertainty, such as emergency response, where predicting the specific impacts of a policy apriori is implausible. Large Language Model (LLM) agent simulations have been proposed as tools to support policymakers under these conditions, yet little is known about how such simulations become useful for real-world policy practice. To address this gap, we conducted a year-long, stakeholder-engaged design process with a university emergency preparedness team. Through iterative design cycles, we developed and refined an LLM agent simulation of a large-scale campus gathering, ultimately scaling to 13,000 agents that modeled crowd movement and communication under various emergency scenarios. Rather than producing predictive forecasts, these simulations supported policy practice by shaping volunteer training, evacuation procedures, and infrastructure planning. Analyzing these findings, we identify three design process implications for making LLM agent simulations that are useful for policy practice: start from verifiable scenarios to bootstrap trust, use preliminary simulations to elicit tacit domain knowledge, and treat simulation capabilities and policy implementation as co-evolving.

cs.HC↗

arXiv2025

Martingale Score: An Unsupervised Metric for Bayesian Rationality in LLM Reasoning

Recent advances in reasoning techniques have substantially improved the performance of large language models (LLMs), raising expectations for their ability to provide accurate, truthful, and reliable information. However, emerging evidence suggests that iterative reasoning may foster belief entrenchment and confirmation bias, rather than enhancing truth-seeking behavior. In this study, we propose a systematic evaluation framework for belief entrenchment in LLM reasoning by leveraging the Martingale property from Bayesian statistics. This property implies that, under rational belief updating, the expected value of future beliefs should remain equal to the current belief, i.e., belief updates are unpredictable from the current belief. We propose the unsupervised, regression-based Martingale Score to measure violations of this property, which signal deviation from the Bayesian ability of updating on new evidence. In open-ended problem domains including event forecasting, value-laden questions, and academic paper review, we find such violations to be widespread across models and setups, where the current belief positively predicts future belief updates, a phenomenon which we term belief entrenchment. We identify the models, reasoning techniques, and domains more prone to belief entrenchment. Finally, we validate the Martingale Score by showing that it predicts ground-truth accuracy on problem domains where ground truth labels are available. This indicates that, while designed as an unsupervised metric that operates even in domains without access to ground truth, the Martingale Score is a useful proxy of the truth-seeking ability of a reasoning process.

cs.AI↗

arXiv2025

AI-Mediated Communication Reshapes Social Structure in Opinion-Diverse Groups

Group segregation or cohesion can emerge from micro-level communication, and AI-assisted messaging may shape this process. Here, we report a preregistered online experiment (N = 557 across 60 sessions) in which participants discussed controversial political topics over multiple rounds and could freely change groups. Some participants received real-time message suggestions from a large language model (LLM), either personalized to their stance (individual assistance) or incorporating their group members' perspectives (relational assistance). We find that small variations in AI-mediated communication cascade into macro-level differences in group composition. Participants with individual assistance send more messages and show greater stance-based clustering, whereas those with relational assistance use more receptive language and form more heterogeneous ties. Hybrid expressive processes-jointly produced by humans and AI-can reshape collective organization. The patterns of structural division and cohesion depend on how AI incorporates users' interaction context.

cs.SI↗

arXiv2025

Spontaneous Giving and Calculated Greed in Language Models

Large language models demonstrate strong problem-solving abilities through reasoning techniques such as chain-of-thought prompting and reflection. However, it remains unclear whether these reasoning capabilities extend to a form of social intelligence: making effective decisions in cooperative contexts. We examine this question using economic games that simulate social dilemmas. First, we apply chain-of-thought and reflection prompting to GPT-4o in a Public Goods Game. We then evaluate multiple off-the-shelf models across six cooperation and punishment games, comparing those with and without explicit reasoning mechanisms. We find that reasoning models consistently reduce cooperation and norm enforcement, favoring individual rationality. In repeated interactions, groups with more reasoning agents exhibit lower collective gains. These behaviors mirror human patterns of "spontaneous giving and calculated greed." Our findings underscore the need for LLM architectures that incorporate social intelligence alongside reasoning, to help address--rather than reinforce--the challenges of collective action.

cs.CL↗

arXiv2025

Actions Speak Louder than Words: Agent Decisions Reveal Implicit Biases in Language Models

While advances in fairness and alignment have helped mitigate overt biases exhibited by large language models (LLMs) when explicitly prompted, we hypothesize that these models may still exhibit implicit biases when simulating human behavior. To test this hypothesis, we propose a technique to systematically uncover such biases across a broad range of sociodemographic categories by assessing decision-making disparities among agents with LLM-generated, sociodemographically-informed personas. Using our technique, we tested six LLMs across three sociodemographic groups and four decision-making scenarios. Our results show that state-of-the-art LLMs exhibit significant sociodemographic disparities in nearly all simulations, with more advanced models exhibiting greater implicit biases despite reducing explicit biases. Furthermore, when comparing our findings to real-world disparities reported in empirical studies, we find that the biases we uncovered are directionally aligned but markedly amplified. This directional alignment highlights the utility of our technique in uncovering systematic biases in LLMs rather than random variations; moreover, the presence and amplification of implicit biases emphasizes the need for novel strategies to address these biases.

cs.CL↗

arXiv2023

Deconstructing Cooperation and Ostracism via Multi-Agent Reinforcement Learning

Cooperation is challenging in biological systems, human societies, and multi-agent systems in general. While a group can benefit when everyone cooperates, it is tempting for each agent to act selfishly instead. Prior human studies show that people can overcome such social dilemmas while choosing interaction partners, i.e., strategic network rewiring. However, little is known about how agents, including humans, can learn about cooperation from strategic rewiring and vice versa. Here, we perform multi-agent reinforcement learning simulations in which two agents play the Prisoner's Dilemma game iteratively. Each agent has two policies: one controls whether to cooperate or defect; the other controls whether to rewire connections with another agent. This setting enables us to disentangle complex causal dynamics between cooperation and network rewiring. We find that network rewiring facilitates mutual cooperation even when one agent always offers cooperation, which is vulnerable to free-riding. We then confirm that the network-rewiring effect is exerted through agents' learning of ostracism, that is, connecting to cooperators and disconnecting from defectors. However, we also find that ostracism alone is not sufficient to make cooperation emerge. Instead, ostracism emerges from the learning of cooperation, and existing cooperation is subsequently reinforced due to the presence of ostracism. Our findings provide insights into the conditions and mechanisms necessary for the emergence of cooperation with network rewiring.

cs.MA↗

arXiv2023

Rethinking Safe Control in the Presence of Self-Seeking Humans

Safe control methods are often intended to behave safely even in worst-case human uncertainties. However, humans may exploit such safety-first systems, which results in greater risk for everyone. Despite their significance, no prior work has investigated and accounted for such factors in safe control. In this paper, we leverage an interaction-based payoff structure from game theory to model humans' short-sighted, self-seeking behaviors and how humans change their strategies toward machines based on prior experience. We integrate such strategic human behaviors into a safe control architecture. As a result, our approach achieves better safety and performance trade-offs when compared to both deterministic worst-case safe control techniques and equilibrium-based stochastic methods. Our findings suggest an urgent need to fundamentally rethink the safe control framework used in human-technology interaction in pursuit of greater safety for all.

cs.HC↗