# Harnessing Disagreement: Detecting Correlated Agreement Blindness in Multi-Agent Triage

arXiv: https://arxiv.org/abs/2607.19899 · HTML: https://mlsanity.com/paper/2607.19899/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 25 references that are in the map.
- First posted on arXiv 22 Jul 2026.

## Builds on (most cited first, 11 of 11)

- [XGBoost: A Scalable Tree Boosting System](https://mlsanity.com/paper/1603.02754/) (2016)
- [Billion-scale similarity search with GPUs](https://mlsanity.com/paper/1702.08734/) (2017)
- [Why Do Multi-Agent LLM Systems Fail?](https://mlsanity.com/paper/2503.13657/) (2025)
- [Multi-Agent Risks from Advanced AI](https://mlsanity.com/paper/2502.14143/) (2025)
- [A Unified Theory of Diversity in Ensemble Learning](https://mlsanity.com/paper/2301.03962/) (2023)
- [Great Models Think Alike and this Undermines AI Oversight](https://mlsanity.com/paper/2502.04313/) (2025)
- [NetFlow Datasets for Machine Learning-based Network Intrusion Detection Systems](https://mlsanity.com/paper/2011.09144/) (2020)
- [Correlated Errors in Large Language Models](https://mlsanity.com/paper/2506.07962/) (2025)
- [Risk Analysis Techniques for Governed LLM-based Multi-Agent Systems](https://mlsanity.com/paper/2508.05687/) (2025)
- [Don't Always Pick the Highest-Performing Model: An Information Theoretic View of LLM Ensemble Selection](https://mlsanity.com/paper/2602.08003/) (2026)
- [A Statistical Framework for Auditing Behavioral Dependence and Induced Bias in LLM Judges](https://mlsanity.com/paper/2604.07650/) (2026)

## Built on by (most cited first, 1 of 1)

- [Agent Behavioral Contracts II: Certifying Compositional Reliability Without Assuming Independence](https://mlsanity.com/paper/2608.12895/) (2026)

## Abstract (arXiv)

Disagreement-triggered escalation can create a structural blind spot in multi-agent arbitration: as base learners improve, they tend to converge, weakening safety monitoring where correlated failures concentrate. We term this correlated agreement blindness and present ARAT (Arbitrated Reasoning Agents for Alarm Triage), a directed-star system combining an inductive Random Forest (RF) agent, an analogical case-based k-nearest neighbour (k-NN) agent, and a calibrated meta-model to mitigate this effect. On 82,332 holdout samples from the UNSW-NB15 network intrusion detection dataset, 57.2% of errors occur under agreement and 90.6% of dangerous under-predictions evade disagreement-based monitoring even after conservative override; ablation shows that strengthening base learners increases error correlation while reducing disagreement. ARAT reduces under-prediction relative to soft voting from 4.80% to 1.70% via conservative override (-2.6pp) and a safety-flag gate (-0.5pp), demonstrating architectural gains. Cross-dataset validation on clinical readmission supports these indicators, suggesting that diversification improves safety only when it generates productive disagreement rather than convergence. These results indicate that disagreement-triggered escalation can be blind to correlated failure, a risk that may intensify as agentic pipelines deploy increasingly capable, correlated models.