DeepMind Embeds a Secret Watermark in Artificial Intelligence Generated Proteins
SynthID Bio is a technique developed by Google DeepMind which places a hidden watermark into artificially generated proteins while maintaining functionality.
Google DeepMind released SynthID Bio — a method that embeds a hidden watermark directly into the amino acid sequence and the 3D structure of an AI-designed protein. They tested it on three proteins, including the receptor-binding domain of the SARS-CoV-2 spike protein: the watermark is detected almost always, while binding strength and natural sequence diversity in the watermarked proteins don't differ from regular ones. The method was also tested on living bacteriophages, and they stayed functional even with the watermark inside.
This has nothing to do with writing code, but the logic is the same as with watermarks in text and images: the more AI creates, the more it matters to know what's AI-made and what isn't. In biology a mistake costs more, so the method is stricter: the watermark holds even through actually synthesizing the protein in a test tube, not just in a file on a computer. For now this is code and data for researchers, not a ready service, so I wouldn't expect to use this in my own work right now.
Source: deepmind.google
Free course
Stop reading about AI — start building with it
The free Claude Code course: your first site, tool or game — no coding. No upsells, no cross-sells — nothing to buy here.
Start free →
Author
Evgenii Arsentev
PhD · Founder & CEO, Dad4Kids
Articles · Latest articles