Skip to content
DraconiticResearch
Search Sources API
203,952 studies31 sourcesUpdated 38 min ago
PreprintCC BY 4.0Zenodo

Classify Before You Embed: Access Control for Retrieval-Augmented AI Over an Organisation's Archive

  • Micky Irons1
  1. 1Mickai LTD
Open at sourcePDFDOICite

Abstract

A practical methodology for document classification and compartmentalised permissioning before an AI system gets retrieval access to an organisation's archive, with the confused-deputy failure mode that results from skipping it.

Related

  1. Integration of Retrieval-Augmented Generation for Knowledge Access in the ELBE Accelerator Control System
    Najmeh Mirian
    arXivPreprintCC BY 4.0Full textApplied physics & instrumentationComputational physicsarXiv
  2. Conformal Factuality Control for Multi-Hop Retrieval-Augmented Generation
    Muhammad Aimal Rehman, Chi-Kuang Yeh
    arXivPreprintarXiv licenceFull textStatisticsarXiv
  3. Beyond Search Scores: Retrieval Quality, Computational Cost, and LLM Judgment Sensitivity in a Multilingual Research Archive
    Oleksandr Pylypenko
    ZenodoPreprintCC BY 4.0Full textZenodo
  4. Graph-Enhanced Retrieval-Augmented Generation for Postoperative Patient Question Answering: A Comparative Evaluation of Knowledge Graph, Retrieval-Augmented Generation, and Hybrid GraphRAG Approaches
    Srinivasagam Prabha, Bernardo G. Collaco, Carina Rosa Malena +3
    Bioengineering (Basel, Switzerland)Peer-reviewedCC BYFull textEurope PMC
  5. GEO-BERT: A Retrieval augmented LLM Framework for Intelligent Dataset Retrieval from the Gene Expression Omnibus
    Maruf Islam Shuvo, Monjura Afrin Rumi, Naznin Akhter +3
    bioRxivPreprintAll rights reservedFull textBioinformatics & computational biologybioRxivPMC

Cite

BibTeXRIS
BibTeX
@misc{irons2026classify,
  title = {{Classify Before You Embed: Access Control for Retrieval-Augmented AI Over an Organisation's Archive}},
  author = {Micky Irons},
  howpublished = {Zenodo},
  year = {2026},
  month = oct,
  doi = {10.5281/zenodo.23513827},
  url = {https://zenodo.org/records/23513827},
  note = {CC BY 4.0},
}
RIS
TY  - UNPB
TI  - Classify Before You Embed: Access Control for Retrieval-Augmented AI Over an Organisation's Archive
AU  - Micky Irons
PY  - 2026
DA  - 2026/10/20
PB  - Zenodo
DO  - 10.5281/zenodo.23513827
UR  - https://zenodo.org/records/23513827
AB  - A practical methodology for document classification and compartmentalised permissioning before an AI system gets retrieval access to an organisation's archive, with the confused-deputy failure mode that results from skipping it.
ER  -