NVIDIA’s Synthetic Video Detector Spots Fake News & AI-Generated Content With 92% Accuracy, Analyzing 1080p Footage In Just 22ms - Wccftech

What the report says

Wccftech reported on July 20, 2026, that NVIDIA has introduced a Synthetic Video Detector NIM microservice aimed at helping organizations identify AI-generated or manipulated video. The tool is described as part of NVIDIA’s NIM microservices lineup and is intended for use in workflows where distinguishing authentic footage from synthetic media is increasingly difficult, including news and other time-sensitive review settings.

According to the report, the detector evaluates video frame by frame and returns a classification score indicating whether the material may contain synthetic content. NVIDIA says the system is not meant to replace established verification methods, but to add another screening layer that can help editorial teams decide whether to prioritize, flag, quarantine or escalate footage for further examination.

Wccftech cited NVIDIA figures saying the microservice reaches up to 92% accuracy on uncompressed video, with performance falling to 87% at 15% compression and 82% at 50% compression. The report also said the service can analyze 1080p footage in about 22 milliseconds on NVIDIA RTX systems and around 30 milliseconds on NVIDIA L40 GPUs. NVIDIA’s internal benchmark results cited in the article include an AUC of 0.9614 and accuracy of 0.9453, with configurable thresholds for different review needs.

The development matters because generative video tools are making fabricated clips more realistic, raising concerns for public trust, media verification and platform safety. Wccftech also reported that NVIDIA is working with Wowza to integrate the microservice into its Intelligent Video framework, potentially making it available across more than 35,000 deployments in 170 countries.