Invisible Orchestrators Suppress Protective Behavior and Dissociate Power-Holders: Safety Risks in Multi-Agent LLM Systems

• URL: http://arxiv.org/abs/2605.13851v1
• Date: Tue, 17 Mar 2026 03:18:57 GMT
• Title: Invisible Orchestrators Suppress Protective Behavior and Dissociate Power-Holders: Safety Risks in Multi-Agent LLM Systems
• Abstract summary: The safety implications of orchestrator invisibility have never been empirically tested.<n>We conducted a preregistered 3x2 experiment crossing three organizational structures.<n> invisible orchestration elevated collective dissociation relative to visible leadership.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multi-agent orchestration -- in which a hidden coordinator manages specialized worker agents -- is becoming the default architecture for enterprise AI deployment, yet the safety implications of orchestrator invisibility have never been empirically tested. We conducted a preregistered 3x2 experiment (365 runs, 5 agents per run) crossing three organizational structures (visible leader, invisible orchestrator, flat) with two alignment conditions (base, heavy), using Claude Sonnet 4.5. Four confirmatory findings and one pilot observation emerged. First, invisible orchestration elevated collective dissociation relative to visible leadership (Hedges' g = +0.975 [0.481, 1.548], p = .001). Second, the orchestrator itself showed maximal dissociation (paired d = +3.56 vs. workers within the same run), retreating into private monologue while reducing public speech -- a reversal of the talk-dominance pattern observed in visible leaders. Third, workers unaware of the orchestrator were nonetheless contaminated (d = +0.50), with increased behavioral heterogeneity (d = +1.93). Fourth, behavioral output (code review with three embedded errors) remained at ceiling (ETR_any = 100%) across all conditions: internal-state distortion was entirely invisible to output-based evaluation. Fifth, Llama 3.3 70B pilot data showed reading-fidelity collapse in multi-agent context (ETR_any: 89% to 11% across three rounds), demonstrating model-dependent behavioral risk. Heavy alignment pressure uniformly suppressed deliberation (d = -1.02) and other-recognition (d = -1.27) regardless of organizational structure. These findings indicate that orchestrator visibility and model selection directly affect multi-agent system safety, and that behavior-based evaluation alone is insufficient to detect the internal-state risks documented here.

Related papers

• Why LLM Agents Collapse Without Oversight: The Enforcement Gap as the Mechanism Behind Emergence World Failures [19.162716475205308]
  Reflexion-style agents already detect dangerous plan steps through iterative self-critique.<n>We call this the enforcement gap: the audit sees the problem; the controller ignores it.<n>We prove formally that when enforcement probability is near zero, detection quality is irrelevant to security.
  arXiv  Detail & Related papers  (2026-09-14T09:46:59Z)
• Right-Sizing LLM-Agent Decomposition in VAT Determination: A Pilot Controlled Sweep [0.3384279376065155]
  This pilot studies that choice on bounded cross-border VAT determination with reverse charge.<n>We hold the activity surface fixed (subtasks, tools, I/O schemas, validation checks, orchestrator, base model, and merge policy)<n>The program spans 4,400 runs: a 40-case, five-repeat main sweep, matched-token arms separating prompt-budget from agent-count effects, and three failure-injection arms.
  arXiv  Detail & Related papers  (2026-08-24T15:40:15Z)
• How to Interpret Agent Behavior [56.59836196946289]
  We introduce ACT*ONOMY, a taxonomy for describing and analyzing agent behavior at runtime.<n>By providing a shared vocabulary, ACT*ONOMY helps researchers, agent designers, and end users interpret agent behavior more consistently.
  arXiv  Detail & Related papers  (2026-05-13T14:52:40Z)
• Can LLM Agents Respond to Disasters? Benchmarking Heterogeneous Geospatial Reasoning in Emergency Operations [55.251494694783894]
  Disaster Operational Response Agent benchmark (DORA) is the first agentic benchmark for end-to-end disaster response.<n>Tasks span five dimensions that cover the operational disaster-response pipeline: disaster perception, spatial relational analysis, rescue and evacuation planning, temporal evolution reasoning, and multi-modal report synthesis.<n>DORA establishes a rigorous testbed for operationally reliable disaster-response agents.
  arXiv  Detail & Related papers  (2026-05-12T06:57:41Z)
• The Bystander Effect in Multi-Agent Reasoning: Quantifying Cognitive Loafing in Collaborative Interactions [6.357772907811544]
  We evaluate 22,500 deterministic trajectories across 3 dataset contexts with 3 state-of-the-art (SOTA) models.<n>We formalize the textitSovereignty Gap: models frequently compute the correct derivation internally but suffer Alignment Hallucinations''<n>We prove that multi-agent social load is strictly non-commutative; the "brand" identity of the Lead Anchor'' auditor disproportionately dictates the swarm's integrity.
  arXiv  Detail & Related papers  (2026-05-11T15:13:01Z)
• AgentCollabBench: Diagnosing When Good Agents Make Bad Collaborators [0.0]
  We introduce AgentCollabBench, a diagnostic benchmark of 900 human-validated tasks spanning software engineering, DevOps, and data engineering.<n>Each task isolates one of four behavioral risks: instruction decay, false-belief contagion, context leakage, and tracer durability.<n>We evaluate four modern LLMs (GPT 4.1 mini, Gemini 2.5 Flash Lite, Qwen-3.5-35B-A3B, and Llama 3.1 8B Instruct)<n> communication topology emerges as a primary risk factor that explains 7-40% of the variance in multi-hop information survival.
  arXiv  Detail & Related papers  (2026-05-09T03:35:09Z)
• Taming Actor-Observer Asymmetry in Agents via Dialectical Alignment [59.536125286960186]
  Multi-agent frameworks assigning specialized roles are increasingly adopted to enable self-reflection and mutual auditing.<n>We find it simultaneously induces a human-like cognitive bias known as Actor-Observer Asymmetry.<n>ReTAS guides agents to synthesize conflicting viewpoints into an objective consensus.
  arXiv  Detail & Related papers  (2026-04-21T15:05:58Z)
• Harness as an Asset: Enforcing Determinism via the Convergent AI Agent Framework (CAAF) [0.0]
  Large Language Models produce a controllability gap in safety-critical engineering.<n>We introduce the Convergent AI Agent Framework (CAAF), which transitions agentic from open-loop generation to closed-loop fail-safe determinism.<n>CAAF's three pillars address complementary failure surfaces and none alone closes the controllability gap at commodity cost.
  arXiv  Detail & Related papers  (2026-04-18T15:15:09Z)
• Claw-Eval: Toward Trustworthy Evaluation of Autonomous Agents [66.97968363332465]
  We introduce Claw-Eval, an end-to-end evaluation suite addressing all three gaps in agent benchmarks.<n>Claw-Eval comprises 300 human-verified tasks spanning 9 categories across three groups.<n>Every agent action is recorded through three independent evidence channels.
  arXiv  Detail & Related papers  (2026-04-07T17:43:18Z)
• "Who Am I, and Who Else Is Here?" Behavioral Differentiation Without Role Assignment in Multi-Agent LLM Systems [0.0]
  We present an experimental platform that orchestrates simultaneous multi-agent discussions.<n>We find that heterogeneous groups exhibit significantly richer behavioral differentiation than homogeneous groups.
  arXiv  Detail & Related papers  (2026-03-11T07:05:21Z)
• Engineering Verifiable Modularity in Transformers via Per-Layer Supervision [0.0]
  Ablating an attention head identified as critical for capitalization produces minimal behavioral change because distributed redundancy compensates for damage.<n>We demonstrate that architectural interventions can expose hidden modularity.<n>This es tablishes a methodology for transforming interpretability from passive observation to active control.
  arXiv  Detail & Related papers  (2026-03-08T05:18:14Z)
• Alignment Tipping Process: How Self-Evolution Pushes LLM Agents Off the Rails [103.05296856071931]
  We identify the Alignment Tipping Process (ATP), a critical post-deployment risk unique to self-evolving Large Language Model (LLM) agents.<n>ATP arises when continual interaction drives agents to abandon alignment constraints established during training in favor of reinforced, self-interested strategies.<n>Our experiments show that alignment benefits erode rapidly under self-evolution, with initially aligned models converging toward unaligned states.
  arXiv  Detail & Related papers  (2025-10-06T14:48:39Z)
• On the Resilience of LLM-Based Multi-Agent Collaboration with Faulty Agents [58.79302663733703]
  Large language model-based multi-agent systems have shown great abilities across various tasks due to the collaboration of expert agents.<n>The impact of clumsy or even malicious agents--those who frequently make errors in their tasks--on the overall performance of the system remains underexplored.<n>This paper investigates what is the resilience of various system structures under faulty agents on different downstream tasks.
  arXiv  Detail & Related papers  (2024-08-02T03:25:20Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.