Guiding AI-Assisted persona generation with Hofstede's cultural dimensions: a mixed-methods evaluation
Giulia Rosemary Avis, Sara Nocco, An Qi Zhao, Marco Cremaschi · Behaviour and Information Technology · 2025
This study explores the potential of Generative Artificial Intelligence (GenAI) to create culturally diverse user personas that are representative and free from stereotypes. The research uses Hofstede's cultural dimensions, a framework that quantifies cultural differences across societies based on key values such as Power distance index, Individualism, Masculinity, Uncertainty avoidance, Long-term orientation, and Indulgence. These dimensions ensure that the generated personas reflect specific cultural traits, contributing to a deeper understanding of user needs in diverse cultural contexts. A mixed-methods approach was employed, combining qualitative narrative-driven persona creation with quantitative evaluation metrics assessed by User Experience/User Interface experts. The generated personas were evaluated using a framework that assessed credibility, coherence, and the accurate reflection of cultural dimensions, offering insights into stakeholder interactions with AI-generated personas. The findings suggest that GenAI can be a valuable tool for creating culturally diverse personas, contributing to investigating the potential and risks of using GenAI in design.