NVIDIA announced its 'Synthetic Video Detector NIM microservice' at SIGGRAPH to help provide trustworthy video content. This service is part of the NVIDIA AI for Media platform and integrates AI-based detection signals into editing and media workflows.

The microservice analyzes video frame-by-frame to generate a classification score indicating the presence of synthetic content. Editing teams can use this score to prioritize videos for review, flag or isolate suspicious footage, and route it for in-depth analysis. This provides additional signals to assist in rapid decision-making without replacing existing verification procedures.

NVIDIA announced that the detector remains functional even after processes such as video compression, resizing, cropping, and re-encoding. Internal testing showed up to 92% accuracy on uncompressed video, 87% accuracy at 15% compression, and 82% accuracy at 50% compression. It processes 1080p video in as little as 22 milliseconds on NVIDIA RTX systems and approximately 30 milliseconds on L40 GPUs.

Media organizations can maintain operational control by deploying the microservice on-premises, at the edge, in hybrid setups, or in approved air-gapped environments where sensitive video is captured or stored. Wowza has integrated this service through the Wowza Video Intelligence Framework, providing real-time detection capabilities across more than 35,000 deployment environments in over 170 countries.