Kakao Adopts Google DeepMind's SynthID to Bolster AI Content Transparency
Translated from Korean and summarized by DistantNews. Read the original for the full story.
At a glance
- Kakao partners with Google DeepMind to enhance transparency of AI-generated content.
- The collaboration will integrate Google DeepMind's SynthID digital watermarking technology into Kakao's AI model, Kanana.
- This move aims to clearly identify AI-generated text, images, audio, and video, making them distinguishable from human-created content.
Kakao is taking a significant step towards fostering trust in the age of artificial intelligence by collaborating with Google DeepMind. The integration of SynthID, Google's advanced digital watermarking technology, into Kakao's proprietary AI model, Kanana, underscores the company's commitment to transparency.
Kakao is joining hands with Google DeepMind to strengthen the transparency of AI-generated content.
This partnership is crucial for distinguishing AI-generated content from human-created material. As AI becomes increasingly sophisticated, the ability to identify its output is paramount for maintaining authenticity and preventing misinformation. Kakao's proactive approach, by adopting SynthID, positions it as a leader in responsible AI development within South Korea.
Through this collaboration, Kakao will apply Google DeepMind's digital watermarking technology, SynthID, to its own AI model, Kanana.
The technology, which embeds imperceptible watermarks into various forms of media, will allow users to verify the origin of content. This is particularly important for Kakao, a dominant platform in Korea, where the implications of AI-generated content range from creative expression to potential societal disruption. By working with Google DeepMind, Kakao is leveraging global expertise to address a critical challenge.
SynthID is difficult for users to recognize and can be applied to images, text, audio, and video.
Originally published by Chosun Ilbo in Korean. Translated, summarized, and contextualized automatically by DistantNews, with a note on how the source frames the story. Not individually reviewed before publishing. How this works.