Abstract
This study examines how generative AI can transform original characters (OCs) into interactive agents that improve character believability in design contexts. We propose a human–AI co-creation workflow that integrates dialogue, image, and context modalities, stabilizing style and behavior through structured prompts and expert feedback. In a mixed-method evaluation with five expert creators and a general audience, the approach increased perceived realism and emotional authenticity, while psychological complexity changed little—indicating the need for longer-horizon design. Dialogue contributed most to credibility, with context reinforcing coherence and images showing limited gains. We distill design principles for cross-modal alignment, style stability, and progressive complexity, offering a practical path to scalable, collaborative character design and more engaging interactive experiences.
Keywords
Generative AI, Original Character, Conversational Agents, Believability, Multimodal Interaction
DOI
https://doi.org/10.21606/drs.2026.688
Citation
Liang, X., Zou, L., Hu, Y., and Luo, S. (2026) Bringing Characters to Vitality: Enhancing Credibility of Original Characters in Narrative Works by making the OC a Generative AI agent, in Simeone, L., Gray, C. M., Verhoeven, A., de Götzen, A., Bakırlıoğlu, Y., Zohar, H., Stead, M., and Buwert, P. (eds.), DRS2026: Edinburgh, 8–12 June, Edinburgh, United Kingdom. https://doi.org/10.21606/drs.2026.688
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
Included in
Bringing Characters to Vitality: Enhancing Credibility of Original Characters in Narrative Works by making the OC a Generative AI agent
This study examines how generative AI can transform original characters (OCs) into interactive agents that improve character believability in design contexts. We propose a human–AI co-creation workflow that integrates dialogue, image, and context modalities, stabilizing style and behavior through structured prompts and expert feedback. In a mixed-method evaluation with five expert creators and a general audience, the approach increased perceived realism and emotional authenticity, while psychological complexity changed little—indicating the need for longer-horizon design. Dialogue contributed most to credibility, with context reinforcing coherence and images showing limited gains. We distill design principles for cross-modal alignment, style stability, and progressive complexity, offering a practical path to scalable, collaborative character design and more engaging interactive experiences.