Metraipolis › archive › dc_feec600ab73e

When AI Investigators Cannot Trust Their Tools: Conflict-Aware Evidence Verification for Cyber Threat Investigation — arXiv:2610.41191

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.41191
License (as archived): CC0-1.0
Archived: 2026-11-02T07:47:24.766Z
SHA-256 of the archived text: 087741b0a90e6580c6efb47d2f250bafa557067a5175e6625818d18676ab2795
Stable URL of this page: https://metraipolis.com/documento/dc_feec600ab73e

How to cite

When AI Investigators Cannot Trust Their Tools: Conflict-Aware Evidence Verification for Cyber Threat Investigation — arXiv:2610.41191 — arXiv (en), https://arxiv.org/abs/2610.41191 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_feec600ab73e (Metraipolis archive dc_feec600ab73e, sha256 087741b0a90e6580c6efb47d2f250bafa557067a5175e6625818d18676ab2795)

BibTeX

@misc{dcfeec600ab73e,
  title = {When AI Investigators Cannot Trust Their Tools: Conflict-Aware Evidence Verification for Cyber Threat Investigation — arXiv:2610.41191},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41191},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_feec600ab73e; sha256 087741b0a90e6580c6efb47d2f250bafa557067a5175e6625818d18676ab2795},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41191. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM agents increasingly triage security alerts by combining the outputs of several detection tools, implicitly treating those tools as trustworthy. Unlike prior work on conflicts between a model's knowledge and its context, we study conflicts among the tools themselves: what happens when this machine-generated evidence is conflicting or wrong. On the corrected CIC-IDS2017 dataset, four heterogeneous detectors emit labels and confidences that we corrupt under a threat model varying the number of corrupted tools, the persistence of the corruption, and the confidence attached to wrong outputs. Across open-weight LLMs from 3B to 14B parameters, unverified investigators are fragile: one corrupted detector out of four lowers naive accuracy to .50-.72, three lower every unverified investigator to .13 or less, and honest disagreement alone makes them abstain on up to 63% of flows. We propose Conflict-Aware Evidence Verification (CAV), built from established components: an equal-weight vote that ignores reported confidences, an evidence-conflict gate, and an independent nearest-neighbour trust check that tests the leading candidates. With its prompt frozen before testing, CAV-LLM improves on the naive investigator in every corrupted condition ($p < 0.001$), reaching .84-.88 with three corrupted detectors against at most .08, while calling the LLM on 0.4% of clean traffic. A fixed verification rule is as accurate or better unless most tools are corrupted. When we make the independent channel itself noisy, CAV inherits its errors: it stays ahead when two or more tools are corrupted, but adding the channel as an extra voter is better when at most one is. Verification thus moves trust from the tools to the check, which must be measured and protected in turn.

Tema degli esploratori: fiducia digitale e crittografia (elenco arXiv cs.CR, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.41191 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-02T07:39:31.053Z dalla pagina https://arxiv.org/list/cs.CR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).