Kakao Becomes First Asian Firm to Adopt Google's SynthID AI Watermark

Kakao (035720) said Thursday it has become the first company in Asia to adopt SynthID, the digital watermarking technology built by Google DeepMind, in a move that deepens the Korean platform giant's push toward a more accountable artificial intelligence ecosystem.

Under a technology partnership with Google DeepMind, Kakao will apply SynthID to its in-house Kanana AI models to help verify whether a given piece of media was machine-generated. The system embeds signals that are imperceptible to the human eye and ear across images, text, audio and video, allowing AI-created content to be traced without degrading its quality.

The watermarking approach has gained traction globally as regulators and platforms search for reliable ways to counter deepfakes and coordinated disinformation. Because the marker survives common edits, it can flag AI origins even after content is cropped, reprocessed or reshared.

Starting in the second half of 2026, SynthID markers will be applied to material generated by Kakao's image model, Kanana Collage, and its video model, Kanana Kinema. The company said its first rollout will target Kanana Template, a feature that turns photos exchanged inside KakaoTalk into short AI-generated clips.

Videos made through Kanana Template already carry a visible Kanana logo, but SynthID adds a hidden layer of provenance that remains detectable even if the clip is later altered. That combination is designed to give users a clearer signal about what they are watching as synthetic media spreads across everyday messaging.

Kim Kyung-hoon, Kakao's AI Safety Leader, said securing transparency and reliability in AI output is becoming more important as the technology enters daily life, adding that the company intends to keep cooperating with Google on responsible AI. Pushmeet Kohli, a vice president at Google DeepMind, said partnerships with leading firms are helping expand SynthID to more platforms and build a more transparent digital ecosystem.

The deal underscores how Kakao is positioning content authenticity and AI governance as central pillars of its expanding Kanana ecosystem, aligning its safety framework with tools already used by one of the world's foremost AI labs.