National cyber-defence organizations must decide when an emerging technology is sufficiently mature to justify further research, experimentation, procurement, or operational deployment. These decisions are difficult because information about technological maturity is fragmented across research papers, patents, standards, prototypes, pilot projects, product announcements, and deployment reports.

Technology maturity is also frequently overstated. Research demonstrations may operate only under controlled conditions, while company announcements may describe planned capabilities as though they were already available. Publication volume and media attention are therefore not reliable indicators of operational readiness.

LLMs could help extract and synthesize evidence about prototypes, performance, scalability, interoperability, standards compliance, adoption barriers, and real-world deployment. Nevertheless, it remains unclear whether LLMs can consistently distinguish conceptual proposals from demonstrated and operational technologies.

This thesis will investigate whether LLMs can provide transparent and reproducible maturity assessments of emerging technologies relevant to national cyber defence.

Objectives

The objective is to design and evaluate an LLM-assisted framework for assessing the maturity, operational readiness, and adoption trajectory of selected cyber-defence technologies.

The framework should produce evidence-based assessments rather than unsupported maturity scores.

Research questions :

• Can LLMs distinguish concepts, laboratory demonstrations, prototypes, pilots, and operational deployments?
• Which technical and non-technical indicators best represent technology maturity?
• How closely do LLM-generated assessments agree with expert assessments?
• Can LLMs distinguish concrete evidence from plans, predictions, and marketing claims?
• How accurately can the models identify barriers related to cost, scalability, regulation, interoperability, skills, and supply chains?
• Can the framework explain why a technology’s maturity assessment changes over time?
• Does RAG improve the accuracy and traceability of maturity assessments?
• How should uncertainty be represented when the available evidence is incomplete or contradictory?

Expected contributions :

The thesis should produce:

• A structured maturity-assessment framework for a selected technology domain.
• A time-stamped dataset of maturity-related evidence.
• An LLM-assisted pipeline for extracting and classifying maturity indicators.
• A comparison between automated and expert assessments.
• Longitudinal maturity profiles of selected technologies.
• Recommendations for using LLM-generated maturity assessments in strategic decision-making.