Collaborative Generative AI for Cyber Mission Planning, Analysis, and Reporting
Name
CollaborativeGenerativeAI_Quiroga.pdf
Size
1.16 MB
Format
Adobe PDF
Checksum (MD5)
c42fb8308acd3992090cd9fb71594872
Author(s)
Quiroga, Aaron
Kepner, Jeremy
Date Issued
August 20, 2026
Abstract
In order to deter adversaries and ensure the security of networks, Cyber Protection Teams (CPTs) are employed to conduct cyber missions in potentially contested environments. These missions require substantial planning, reporting, and analysis, with operators producing mission artifacts such as crew logs, notes, findings, and observations throughout the mission lifecycle. These artifacts must be reviewed, organized, and consolidated into formal mission products, creating a documentintensive workflow that can reduce the time available for deeper analysis and execution. As cyber threats continue to rapidly evolve their Tactics, Techniques, and Procedures (TTPs), so too must cyber defenders. The work in this paper presents a collaborative architecture that utilizes a Retrieval-Augmented Generation (RAG) pipeline for document creation. Through this pipeline, the architecture ingests mission documents and artifacts, retrieves relevant mission context, and generates reviewable draft updates to support operators and mission leads in creating planning and reporting documents. By combining evidencegrounded retrieval, local model inference, and human review, the proposed architecture aims to provide operational flexibility for a more holistic approach to cyber operations while preserving mission isolation and accountability.
Subjects
cyber protection team
defensive cyber operations
retrieval-augmented generation
MIT Department
Lincoln Laboratory
Terms of Use
All rights reserved
Persistent DSpace Link
