Summary:
- This project investigates how assistive AI can improve the teaching and practice of cyber risk assessment.
- It focuses on developing adaptive cyber risk assessment scenarios using Large Language Models within cyber range environments.
- This knowledge will support more personalised, scalable and realistic cyber security training for learners at different levels.
- The student will design an AI-supported framework, develop a quality assurance mechanism and evaluate the impact of adaptive scenarios on learner performance, engagement and skill transfer.
- Stage 1 Deadline: 29 May 2026
Background:
Cyber risk assessment is a vital skill for organisations,
Traditional training methods often rely on static case studies that are difficult to customise, slow to update, and limited in their ability to reflect the complexity of real organisational environments. This project proposes the use of assistive artificial intelligence, particularly Large Language Models (LLMs), to generate adaptive training scenarios using cyber range that can be tailored to learner needs while maintaining realism, consistency, and pedagogical value.
The researcher will develop and test an AI-supported framework, including a quality assurance mechanism to ensure scenarios are accurate, coherent and trustworthy. They will also evaluate whether adaptive AI-generated scenarios improve learner performance, engagement and transfer of skills compared with traditional training methods. Through this work, the researcher will contribute new knowledge and practical design guidance for scalable, responsible and effective cyber risk assessment training.
The main objectives of the project are:-
To design an assistive AI framework for generating adaptive cyber risk assessment scenarios.
- To develop a quality assurance mechanism that validates the correctness and consistency of generated scenarios.
- To compare the effectiveness of AI-generated adaptive scenarios with traditional case-study-based training.
- To identify which scenario characteristics best support learning and skill development.- To produce practical design guidelines for scalable, trustworthy AI-supported cyber risk assessment training.
Estimated thesis submission: