Project Description

ARVIS: AI-Powered Digital Human

The VAR Lab is developing ARVIS, an AI-powered digital human designed to support communities in rural areas where internet access may be inconsistent or limited.

ARVIS combines a locally hosted large language model with a realistic digital human interface, giving users a face to talk to while accessing information and support. The project uses voice cloning technology from ElevenLabs and a MetaHuman created from a high-resolution Artec 3D scan to create a more natural and approachable interaction.

By combining generative AI, digital doubles, local LLMs, 3D scanning, and voice technology, ARVIS explores how AI systems can become more accessible, human-centered, and useful in places where cloud-based tools may not always be reliable.