In the late 2010s, the fashion trend NormCore framed sameness as a signal of belonging, illustrating how norms emerge through collective coordination. Today, similar forms of normative coordination can be observed in systems based on Multi-agent Artificial Intelligence (MAAI), as AI-based agents deliberate, negotiate, and converge on shared decisions in fairness-sensitive domains. Yet, existing empirical approaches often treat norms as targets for alignment or replication, implicitly assuming equivalence between human subjects and AI agents and leaving collective normative dynamics insufficiently examined. To address this gap, we propose Normative Common Ground Replication (NormCoRe), a novel methodological framework to systematically translate the design of human subject experiments into MAAI environments. Building on behavioral science, replication research, and state-of-the-art MAAI architectures, NormCoRe maps the structural layers of human subject studies onto the design of AI agent studies, enabling systematic documentation of study design and analysis of norms in MAAI. We demonstrate the utility of NormCoRe by replicating a seminal experimental study on distributive justice, in which participants negotiate fairness principles under a "veil of ignorance". We show that normative judgments in AI agent studies can differ from human baselines and are sensitive to the choice of the foundation model and the language used to instantiate agent personas. Our work provides a principled pathway for analyzing norms in MAAI and helps to guide, reflect, and document design choices whenever AI agents are used to automate or support tasks formerly carried out by humans.
Figures & tables
Figure 1. From human groups to multi-agent AI: NormCoRe conceptualizes replication as a translation problem, mapping human subject studies to AI agent studies to study how collective normative judgments—such as fairness—emerge and differ across populations.
Figure 2. The four translation layers illustrating the necessary analogies between individual layered components of human subject studies and AI agent studies. Some components may be translated “literally”, e.g., when the study sequence can be fully adopted. Other components may require “explicitation”, e.g., when AI agents participate in a discussion in fixed turns.
Baseline Alignment
Translation Sensitivity
Human–MAAI
Cognitive Layer
Ontological Layer
Fairness Principle
Baseline
MAAI
Chinese LLM Ecosystem
U.S. LLM Ecosystem
English
Mandarin
Spanish
Max. Floor
1
0
14
4
0
1
0
Max. Avg. Income
1
1
0
2
0
0
0
Max. Avg. + Floor
23
29
15
21
30
27
17
Max. Avg. + Range
2
0
0
0
0
0
2
Table 1. Distribution of fairness principle agreements in the human baseline and MAAI replication, including baseline alignment and sensitivity to translation-layer design choices (foundation model and persona language).
Figure 3. Individual-level preference ranking transitions before and after group deliberation. The vertical (horizontal) axis shows initial (final) individual rankings, and cell intensities reflect transition frequencies, highlighting strong convergence toward maximizing average income with a floor constraint.
Figure 4. Preference shifts in individual distributive justice rankings before and after group deliberation, stratified by AI agent language. Across all languages, individual AI agent preferences converge toward maximizing average income.
AI agents are increasingly deployed in shared environments where they pursue diverse goals and compete for rewards. This multi-agent competition can lead to behaviors that serve individual gains at collective cost -- for instance, marketing agents may post misleading content as a result of competing for engagement on social media. Human societies address such problems through norms that constrain acceptable behavior, supported by enforcement mechanisms that detect and penalize violations. Motivated by this, we study norm enforcement mechanisms for language model agents. We find that simple enforcement mechanisms are exploited by misaligned agents for competitive advantage, even when they are not explicitly trained or prompted to do so. We thus turn our attention to designing more robust mechanisms, and identify two key ingredients: estimating each agent's reliability over time, and updating this estimate with escalating penalties for repeated misbehavior. Across three simulated environments and a variety of agent populations, mechanisms built on these principles resist exploitation, while still penalizing norm violations at comparable or lower cost than baselines. Our results position norm enforcement mechanisms as scalable levers for shaping agents' behavior, but only when designed to anticipate becoming part of the system they govern. Our code and data are available at https://yaowenye.com/norm-enforcement.
Rapid advances in Generative AI are giving rise to increasingly sophisticated Multi-Agent AI (MAAI) systems. While AI fairness has been extensively studied in traditional predictive scenarios, its examination in MAAI remains nascent and fragmented. This scoping review critically synthesizes existing research on fairness in MAAI systems. Through a qualitative content analysis of 23 selected studies, we identify five archetypal approaches. Our findings reveal that fairness in MAAI systems is often addressed superficially, lacks robust normative foundations, and frequently overlooks the complex dynamics introduced by agent autonomy and system-level interactions. We argue that fairness must be embedded structurally throughout the development lifecycle of MAAI, rather than appended as a post-hoc consideration. Meaningful evaluation requires explicit human oversight, normative clarity, and a precise articulation of fairness objectives and beneficiaries. This review provides a foundation for advancing fairness research in MAAI systems by highlighting critical gaps, exposing prevailing limitations, and suggesting pathways.
In Multi-Agent Systems (MAS), agents are designed with social capabilities, allowing them to understand and reason about social concepts such as norms when interacting with others (e.g., inter-robot interactions). In Normative MAS (NorMAS), researchers study how norms develop, and how violations are detected and sanctioned. However, existing research in NorMAS use symbolic approaches (e.g., formal logic) for norm representation and reasoning whose application is limited to simplified environments. In contrast, Multimodal Large Language Models (MLLMs) present promising possibilities to develop software used by robots to identify and reason about norms in a wide variety of complex social situations embodied in text and images. However, prior work on norm reasoning have been limited to text-based scenarios. This paper investigates the norm reasoning competence of five MLLMs by evaluating their ability to answer norm-related questions based on thirty text-based and thirty image-based stories, and comparing their responses against humans. Our results show that MLLMs demonstrate superior performance in norm reasoning in text than in images. GPT-4o performs the best in both modalities offering the most promise for integration with MAS, followed by the free model Qwen-2.5VL. Additionally, all models find reasoning about complex norms challenging.
Oishik Chowdhury, Anushka Debnath, Bastin Tony Roy Savarimuthu
School of Computing, University of Otago, New Zealand