The AI So-Go-Chi Center is pleased to announce that the following paper, co-authored by Associate Professor Teeradaj Racharak and Assistant Professor Watanee Jearanaiwongkul, has been accepted for presentation at the 15th International Joint Conference on Knowledge Graphs (IJCKG 2026). The conference will be held in Bangkok, Thailand, from November 19 to 21, 2026. IJCKG is an academic forum that brings together the international knowledge graph community and related areas, such as the Semantic Web, Linked Data, and AI on the Web, to present innovative research results and novel applications.

Accepted Paper

Chanphot Wongtaweesup, Teeradaj Racharak, Watanee Jearanaiwongkul, Chutiporn Anutariya, Frederic Andres,

Ontology-Guided Scenario Control Generation for GenAI in a Cybersecurity Serious Game

In Proceedings of the 15th International Joint Conference on Knowledge Graphs (IJCKG 2026), Bangkok, Thailand, November 19–21, 2026.

Generative AI (GenAI) can expand the pool of scenarios used in cybersecurity education, but fluent outputs may still omit required controls or provenance. Focusing on PersonaShield Quest, a cybersecurity serious game covering phishing, smishing, vishing, and legitimate communication, this paper proposes the PersonaShield Quest Scenario Control Ontology (PSQ-SCO), an OWL/RDF ontology that serves as a pre-generation contract for scenario content.

PSQ-SCO is built around four competency questions that specify learner context, cybersecurity structure, validation rules, and evidence, and it is integrated with the game engine. Comparing four generation modes (prompt only, text-based retrieval-augmented generation (RAG), structured ontology prompting, and ontology-to-text RAG) across three local language models, the study shows that the ontology adds auditable controls and provenance even when baseline prompting already satisfies the game’s format requirements.

The work offers a practical, knowledge-graph-based framework for making GenAI-generated educational scenarios more reliable and transparent.

Congratulations to the authors on this achievement!