# The Failure Happens Before the Drift: The Social Dynamics of Values in LLM Agent Societies

arXiv: https://arxiv.org/abs/2609.05514 · HTML: https://mlsanity.com/paper/2609.05514/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use.
- Builds on 18 of its 48 references that are in the map.
- First posted on arXiv 31 Aug 2026.

## Placed in

- LLM agents and tool use / [Social simulation and multi-agent debate](https://mlsanity.com/agents/social-simulation-and-multi-agent-debate/)

## Builds on (most cited first, 18 of 18)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Generative Agents: Interactive Simulacra of Human Behavior](https://mlsanity.com/paper/2304.03442/) (2023)
- [MemGPT: Towards LLMs as Operating Systems](https://mlsanity.com/paper/2310.08560/) (2023)
- [Whose Opinions Do Language Models Reflect?](https://mlsanity.com/paper/2303.17548/) (2023)
- [SOTOPIA: Interactive Evaluation for Social Intelligence in Language Agents](https://mlsanity.com/paper/2310.11667/) (2023)
- [Assessing Cross-Cultural Alignment between ChatGPT and Human Societies: An Empirical Study](https://mlsanity.com/paper/2303.17466/) (2023)
- [Do LLMs exhibit human-like response biases? A case study in survey design](https://mlsanity.com/paper/2311.04076/) (2023)
- [Winning Arguments: Interaction Dynamics and Persuasion Strategies in Good-faith Online Discussions](https://mlsanity.com/paper/1602.01103/) (2016)
- [A Computational Approach to Politeness with Application to Social Factors](https://mlsanity.com/paper/1306.6078/) (2013)
- [The Ghost in the Machine has an American accent: value conflict in GPT-3](https://mlsanity.com/paper/2203.07785/) (2022)
- [Mind the Value-Action Gap: Do LLMs Act in Alignment with Their Values?](https://mlsanity.com/paper/2501.15463/) (2025)
- [EtiCor: Corpus for Analyzing LLMs for Etiquettes](https://mlsanity.com/paper/2310.18974/) (2023)
- [The Role of Pragmatic and Discourse Context in Determining Argument Impact](https://mlsanity.com/paper/2004.03034/) (2020)
- [Opening up Minds with Argumentative Dialogues](https://mlsanity.com/paper/2301.06400/) (2023)
- [Proto: A Neural Cocktail for Generating Appealing Conversations](https://mlsanity.com/paper/2109.02513/) (2021)
- [Overstating Attitudes, Ignoring Networks: LLM Biases in Simulating Misinformation Susceptibility](https://mlsanity.com/paper/2602.04674/) (2026)
- [ValueFlow: Measuring the Propagation of Value Perturbations in Multi-Agent LLM Systems](https://mlsanity.com/paper/2602.08567/) (2026)
- [User Behavior Prediction as a Generic, Robust, Scalable, and Low-Cost Evaluation Strategy for Estimating Generalization in LLMs](https://mlsanity.com/paper/2507.05266/) (2025)

## Abstract (arXiv)

Large Language Model (LLM)-based agents are increasingly used as proxies for human participants in social science research, yet it remains unclear whether they can faithfully simulate diverse and conflicting human value systems. We present a World Values Survey (WVS)-grounded simulation framework where culturally diverse agents with different communication styles engage in longitudinal, value-laden discussions. Across approximately 4,000 conversations involving 1,200 personas, 15 topics, and three models (GPT-4o, Gemini-2.5-Flash, and Gemma-4-E4B), we evaluate value faithfulness, value drift, and conversational realism. We find that more than 50% of personas fail to express their assigned WVS profiles from the outset, while 2-7% drift after repeated conversations. Ablations removing demographic details improve faithfulness for some models but do not change the broader trend: simulated value distributions still systematically deviate from the assigned WVS profiles. Compared to human discussions, simulated dialogues show a different trade-off between stylistic consistency and semantic diversity, often producing content-wise varied but stylistically repetitive exchanges. These findings suggest that current LLM agents can generate plausible conversations, but remain limited proxies for representing and preserving diverse human value profiles over time.