Skip to content
UCLA Develops Optical Chip for Deepfake Detection with 97.79% Accuracy

UCLA Develops Optical Chip for Deepfake Detection with 97.79% Accuracy

First seen 4 Oct 2026, 04:08 UTC • •
API

Article Content

Browse articles
ThreatCluster AI
ThreatCluster •October 4, 2026 at 06:06 UTC
  • •UCLA's optical-neural processor detects deepfakes with 97.79% accuracy across 15 streams.
  • •The system features 99.86% sensitivity and 95.72% specificity, making it highly effective.
  • •This technology offers a scalable solution to the increasing challenges posed by deepfake videos.

Researchers at UCLA have created an optical-neural processor that detects deepfake videos with an average accuracy of 97.79% across 15 video streams simultaneously. The system achieves a sensitivity of 99.86% and specificity of 95.72%, making it a promising first layer of defense against manipulated content. The technology utilizes light to process video frames through diffractive layers, allowing for rapid analysis without relying on traditional GPU processing. In tests, the chip maintained 96.13% accuracy when screening 18 videos and achieved 94.80% accuracy against Google's VEO-3 generated videos. This innovation addresses the growing challenge of deepfake detection as generative AI improves, providing a scalable and energy-efficient solution. The study detailing this technology was published in *eLight* and is credited to a team led by Professor Aydogan Ozcan.

Start a free Starter trial for enhanced analysis

Ask AI about this cluster

Updated just now How this analysis works

Timeline

2026-09-29
Study published in eLight
UCLA researchers published their findings on the optical-neural processor for deepfake detection.
ScienceDaily
2026-10-01
ScienceDaily reports on UCLA's technology
ScienceDaily detailed the capabilities and accuracy of UCLA's new optical-neural processor for detecting deepfakes.
ScienceDaily
2026-10-02
Aiweekly.Co covers UCLA's optical chip
Aiweekly.Co reported on the detection accuracy and features of UCLA's optical-neural processor.
Aiweekly.Co

More articles in this cluster (2)

Extracted Entities

1 / 2

Common questions

How does the optical-neural processor work?
It analyzes video frames using light propagation through diffractive layers instead of traditional digital processing.
What is the significance of the detection accuracy?
With 97.79% accuracy and high sensitivity, it serves as an effective first layer of defense against deepfake videos.
Can this technology replace existing detection methods?
It is designed to complement existing digital detectors, not replace them, by providing an initial screening layer.