1 free read left. Create a free account for unlimited reading + AI research.
Create free accountWhitepaper
DiGRA
The Vulnerable Avatars Framework identifies ten contexts of vulnerability and maps them onto aesthetic, narrative, and systemic layers of game design to produce emotionally credible virtual victims.
This paper describes the design process of the Victim Interview Simulator (VIS), a serious game developed within the EU-funded ISEDA project to train police officers in first-contact interviews with victims of domestic abuse. Its central contribution is the Vulnerable Avatars Framework, a structured method for designing emotionally credible virtual victims. The framework identifies ten contexts of vulnerability, including social and psychological factors, types of oppression, and roles in violent situations, and maps them onto aesthetic, narrative, and systemic layers of game design. It also specifies the contributions of social actors such as police officers, psychologists, and victim advocates, alongside development roles including designers, animators, and actors.
The simulator applies this framework through three scenarios depicting women of different ages, backgrounds, and forms of abuse, co-designed with European police agencies, victim-support organizations, and academic partners. Dialogue is driven by a stack-based system rather than rigid branching trees, and trust and impatience variables regulate access to sensitive topics such as sexual violence and coercive control. Technically, the project combines MetaHuman characters with low-cost markerless facial motion capture, recorded with actresses across multiple countries, to convey subtle nonverbal cues of guardedness and hesitation.
A preliminary evaluation used the Perceived Interpersonal Vulnerability (PIV) scale, a measure developed for the project, with 19 practitioners at an ISEDA closing event. Participants consistently perceived the avatar as vulnerable, with emotional reactivity receiving the highest ratings and nonverbal cues reliably detected. Results for negative affect and support-seeking were mixed, which the authors attribute to translation problems and single-item measurement. The authors acknowledge limitations including a small sample, an uncontrolled setting, and linguistic mismatch, and conclude that the framework is a feasible, ethically grounded approach, while stating that measurable effects on professional practice remain to be established.
European Conference on Games Based Learning (ECGBL) · 2024