Google announced at I/O 2026 that OpenAI, Kakao, and ElevenLabs will adopt SynthID, its invisible watermark technology for AI-generated content. This is the first time direct competitors have used a transparency tool developed by the company. SynthID embeds an imperceptible digital signal into images, videos, or audio, resistant to filters and compression. In three years, it has marked over 100 billion visual files and over 60,000 years of audio.
Digital marking that survives everything 🛡️
SynthID works by inserting a digital fingerprint directly into the pixels or sound waves of AI-generated content. This signal is imperceptible to the human eye and ear, but remains detectable even after applying edits such as filters, format changes, or intense compression. The system does not require a centralized database for verification, allowing any user or platform to check the origin of the content locally. Google assures that the process does not degrade the quality of the original material.
The digital bear hug nobody asked for 🐻
So now OpenAI and company are going to use Google's watermark. It's like Coca-Cola lending you its secret formula to use in your homemade cola. The irony is that these same competitors have spent years trying to differentiate themselves from the Mountain View giant, and now they're signing up for its transparency system as if it were a free antivirus pack. At least, when your favorite deepfake is detected, you'll know who to thank for the snitch.