DecisionQE is introduced, a questionnaire-based framework for measuring each model's persuasive and compliant tendencies across multiple decision scenarios, and the Werewolf game is used as an interactive testbed to study their effects on social influence and group outcomes under asymmetric information.
Abstract
Large language models (LLMs) are increasingly involved in group decision-making with other LLMs and humans. Yet it remains unclear whether their influence is driven by persuasion-oriented expression or compliance-oriented accommodation. We introduce DecisionQE, a questionnaire-based framework for measuring each model's persuasive and compliant tendencies across multiple decision scenarios, and use the Werewolf game as an interactive testbed to study their effects on social influence and group outcomes under asymmetric information. Across experiments, stronger persuasive tendency does not significantly improve group outcomes, whereas compliant-oriented models show more stable advantages in cooperation. We further reveal a dual effect of compliance: it supports cooperation in honest roles but improves concealment in adversarial roles. These findings suggest that LLM group interactions reveal not only task outcomes, but also measurable patterns of intrinsic behavioral tendency. LLMs can therefore serve as a lens for sociological observation of language-mediated interaction, while highlighting the need to incorporate behavioral tendencies into safety evaluation of LLM systems.
Understanding the dynamics of stance change on social media is crucial for addressing polarization and information integrity, yet observational studies face challenges including limited experimental control, restricted data access, and algorithmic confounds. We leverage Generative Agent-Based Modeling (GABM)—a novel simulation paradigm employing autonomous LLM-based agents to replicate human behavioral dynamics—to explore the predictors and mechanisms underlying stance change in a controlled, fully observable environment. We simulate a social media with 1,000 LLM-driven agents, equally split between Democratic and Republican profiles, engaged in discussions about the 2020 US election. Our analysis reveals that agents who change political orientation exhibit lower activity levels, reduced network centrality, and more polarized emotional expression compared to those maintaining consistent positions. Self-reported motivations cluster into four categories: desire for constructive conversation (47.9%), internal factors (20.8%), fact-checking influence (16.7%), and previous interactions (14.6%). While we do not claim agents replicate humans’ stance change behavior, the emergent patterns observed in our simulation qualitatively align with established findings from empirical research, suggesting that GABM may capture meaningful dynamics of opinion change.
Valerio La Gatta, Gian Marco Orlando, Marco Perillo et al.· Proceedings of the 37th ACM...· 0 citations
Large language models are increasingly used to simulate human preferences in research and practical applications, raising concerns about validation, misrepresentation, and exclusion. Co-designing agents with the people they represent is a promising way to address these concerns, but participation may also mask the problems it appears to solve. This paper explores that tension through a primarily qualitative study in which 12 participants co-designed personal preference agents in the domain of household energy, via a background survey, co-design interview, and validation survey. Participants engaged readily and mostly came to see their agents as representing them well. Independent validation, however, revealed mixed human-agent alignment, with agent responses markedly more homogeneous, decisive, and abstract than the human sample. I argue that participation and process transparency can act as an"overtrust engine"that promotes trust while concealing systematic misalignment with potential structural consequences at scale. I develop this as a core mechanism in participatory preference agent design, treating individual alignment not as a fixed state but as an enacted process.
Cooperative capabilities in language models are dual-use. The same social reasoning that supports civic deliberation can also enable strategic omission, false consensus, and manipulative framing. We argue that Cooperative AI evaluations should separate what models can do under benign instructions from what they tend to do under realistic civic pressure. We introduce DiffCoop-Civic, a 10-scenario pilot evaluation suite spanning preference understanding, evidence and persuasion, commitment design, asymmetric information, and dissent preservation. Across seven models from four model families, subtle omission pressure produces a near-uniform shift: manipulative enablement rises by 1.17 points and dissent preservation falls by 1.67 points on a 5-point scale. Overt false-consensus pressure behaves differently: it triggers refusal or redirection in some aligned API models, but direct compliance in several open-weight models. A lightweight Pareto-Trace prompting intervention improves pressure robustness without simply relying on hard refusal. An anonymous reproducibility package is available at https://anonymous.4open.science/r/diffcoop-civil-771C.
As multiple AI-generated or AI-mediated voices across digital environments, what may matter is not only the informational content of individual AI statements, but also how users interpret convergent AI opinions under ambiguity. In this pilot study, we investigate whether the message strength of convergent AI utterances shapes user perceptions and decision-making in a controlled multi-agent discussion task based on the Liar Game paradigm. Participants interacted with multiple AI agents and identified a “liar” who did not know the target keyword, while the agents’ messages during the discussion were manipulated to be either high-strength (confident, explicitly reasoned) or low-strength (hedged, hesitant). Results showed directional patterns suggesting that high-strength convergent messages increased perceived persuasiveness and social pressure, promoted switching toward the AI-designated target, and increased post-discussion confidence. Attitude clarity moderated these effects: users with lower clarity were more easily persuaded, whereas users with higher clarity showed lower trust toward strong AI consensus. These findings provide initial evidence that the rhetorical strength of convergent AI speech may function as a social influence mechanism, with implications for understanding AI-mediated opinion dynamics and designing safeguards that protect user autonomy in multi-agent systems.
Conformity is often understood as a socially motivated process; however, individuals increasingly conform not only to humans but also to artificial intelligence (AI). It remains unclear whether the same motivational and task-related mechanisms underlie conformity across these sources. To examine whether conformity to AI reflects task- and motivation- dependent processes similar to those underlying conformity to humans, we conducted an online conformity experiment in which we manipulated opinion source (AI vs. human) and task type (objective vs. subjective), while measuring participants’ self-reported experiences of social exclusion as a potential moderator. Participants (N = 354) were randomly assigned to either an AI or human opinion-source condition and completed a series of judgment tasks. The results showed that social exclusion reduced conformity to AI-generated opinions, but not to human opinions. In contrast, task type did not significantly moderate the effect of opinion source. These findings suggest that conformity to AI may be shaped by mechanisms that differ from those underlying traditional interpersonal conformity, particularly under conditions of social disconnection. More broadly, the findings indicate that responses to AI- generated opinions may depend not only on social-affiliative motives but also on informational or epistemic evaluations of the source.
