NAVI advances Generative AI (GenAI) and Virtual Reality (VR) to enable high-performance and hyper-personalised immersive experiences in Virtual Worlds. At its core is a new generation of smart, people-centred and accessible 3D digital assistants powered by GenAI. The project drives innovation across four pillars:
- multimodal and personalised interaction in Virtual Worlds, including sign language support;
- high-performance and safe GenAI at the edge;
- explainable AI and behavioural decision support systems;
- expressive and culturally sensitive 3D avatars.
The resulting assistants will communicate effectively and inclusively with diverse users, provide accurate and reliable knowledge, and deliver personalised, transparent recommendations that support decision making. NAVI involves Social Sciences and Humanities disciplines and the active participation of representatives from European deaf and autistic communities, and collaborates with the AI, Data and Robotics (ADRA) Partnership and the Virtual Worlds Partnership. The assistants will be demonstrated in four VR environments across culture, healthcare, manufacturing and education, reaching TRL 6 during the project. Beyond the project, a clear exploitation plan and dedicated replicability studies will support commercialisation and adoption across the public and private sectors in Europe.
ZSI brings the social sciences perspective on accessibility, inclusion and cultural sensitivity to NAVI. ZSI leads the work on cultural and identity-aware instruction modelling, developing recommendations on social and cultural norms so that GenAI responses fit different user profiles. Throughout development, ZSI aligns and evaluates the NAVI system for accessibility and inclusion, aiming for Societal Readiness Level 6. ZSI also leads the recruitment of a diverse, representative group of participants for the four pilots, and is responsible for two public reports, Barriers Mitigation Guidelines for VR Interaction and NAVI System Accessibility and Inclusion Evaluation.
