Rabawave said on Thursday it jointly developed deepfake pre-emptive prevention technology with the KAIST Cybersecurity Research Center.

The company said the technology goes beyond existing post-detection methods. It is a pre-emptive defence solution that blocks deepfake generation at the source. It can be used not only by global smartphone makers, social media platforms and cloud service companies but also by governments and public institutions responsible for digital safety.

The company stressed it can block deepfake crimes because it causes serious errors when generative AI learns from or processes images protected with fine noise. It said the result is output that is severely distorted or completely loses its function as content instead of producing normal results.

Rabawave explained that it is designed to add fine noise to the original image at a level imperceptible to the human eye. It said ordinary users feel no difference when viewing the photo, but the noise acts as a fatal source of confusion for AI models that generate deepfakes.

Rabawave CEO Junyeop Kim (김준엽) said deepfake crimes are a global social problem and that fundamental prevention and problem-solving are impossible without a technological response. He said he expects the company to contribute to creating a safer digital environment through cooperation with global platforms and device manufacturers.