mlsanityData to 7 Oct 2026

Shared and structured inputs undermine collective random choice by reasoning AI agents

arXiv 2610.09667 ↗First posted 7 Oct 2026cs.AI physics.soc-phcited by 0

On the map: LLM agents and tool use → Social simulation and multi-agent debate

Abstract (arXiv)

Random selection is widely used in resource allocation and auditing, making reliable implementation essential for AI-agent systems. Behavioural tests across six reasoning models uncovered threshold and divisibility rules used in identifier-based choices. For threshold-following GPT-6 Sol and Gemini 3.8 Flash, single-agent measurements prospectively predicted correlated participation under shared identifiers and biased participation under distinct identifiers with common timestamp bits. Changing dates, formats and identifier labels revealed when these predictions held. Explicit instructions to randomize independently reduced but did not eliminate shared-input correlation. To test implications for oversight, we asked four models to select customer requests randomly for human review. GPT-6 Sol approached the target rate while selecting predictably from identifiers; the others rarely selected requests. All four closely followed supplied random draws. These findings expose collective and audit vulnerabilities that selection rates alone miss, making input-dependent bias, correlation and predictability central targets for agent evaluation.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use. Builds on 11 of its 37 references that are in the map. First posted on arXiv 7 Oct 2026.

Builds on

11 of 37 references.

Closest ancestors · 2

Same direction, newest first

  1. 2026Do LLMs Beat Nash? Testing Decentralized Coordination in Self-Play Multi-Agent Games12 Aug
  2. 2025Emergent Social Dynamics of LLM Agents in the El Farol Bar Problem4 Sep

Other references · 9

  1. 2025Reasoning Models Don't Always Say What They Thinkcited by 388
  2. 2024Heads, Tails, and AI Fails: LLMs, Randomness, and Human Judgmentscited by 23
  3. 2024Subversion Strategy Eval: Can language models statelessly strategize to subvert control protocols?cited by 14
  4. 2025Playing games with Large language models: Randomness and strategycited by 7
  5. 2026The Illusion of Stochasticity in LLMscited by 7
  6. 2026Strategic Algorithmic Monoculture: Experimental Evidence from Coordination Gamescited by 6
  7. 2026Tacit Coordination of Large Language Modelscited by 5
  8. 2025Subversion via Focal Points: Investigating Collusion in LLM Monitoringcited by 3
  9. 2026Frontier Models Can Take Actions at Low Probabilitiescited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026State-dependent error correlations shape voting thresholds in committees of AI agents
  2. 2025A Framework for Studying AI Agent Behavior: Evidence from Consumer Choice Experiments
  3. 2025Human aversion? Do AI Agents Judge Identity More Harshly Than Performance
  4. 2026Epistemic Filtering and Collective Hallucination: A Jury Theorem for Confidence-Calibrated Agents
  5. 2026Language Models Exhibit Inconsistent Biases Towards Algorithmic Agents and Human Experts
  6. 2026Preregistration for Experiments with AI Agents
  7. 2026Aligned Alone, Misaligned Together: Forecasting Adversarial Capture in LLM Agent Populations
  8. 2026AI Coding Agents in Social Science: Methodologically Diverse, Empirically Consistent, Interpretively Vulnerable