Google DeepMind releases SynthID Bio protein watermarking technology

On September 30, 2026, Google DeepMind launched SynthID Bio through official channels (first-hand information), bringing watermarking technology into the field of synthetic biology. The technology embeds detectable signals directly into biological code, adding verifiable watermarks to AI-generated proteins and their structures, and these watermarks can be verified on physically synthesized proteins. Google DeepMind had previously applied SynthID watermarking to text, images, and other content; the launch of SynthID Bio marks the first extension of its watermarking system into synthetic biology, used to identify and verify AI-generated proteins. Later that same day, according to Ars Technica, the Google DeepMind team published a paper proposing a protein watermarking scheme, SynthIDBio, a system that can add watermarks to AI-designed protein sequences without compromising protein function. Built on SynthID technology and integrated into the popular protein design tool ProteinMPNN, the system introduces watermark-consistent amino acids only at positions where function remains viable as it places each amino acid.

AI overview of reporting · Updated 17 days ago