Function-preserving Watermarking of AI-generated Proteins

David Stutz^1,2, Alexander I. Cowen-Rivers^1,2, Guillermo Ortiz-Jimenez¹, Jeremy Ratcliff¹, Vinicius Zambaldi¹, Lindsay Willmore¹, Josh Abramson¹, Harshnira Patani¹, Christina Kouridi¹, Florian Stimberg¹, Mel Vecerik¹, Alex Chu¹, Sukhdeep Singh¹, Sumanth Dathathri¹, Eliseo Papa¹, Valentin De Bortoli¹, Arnaud Doucet¹, Demis Hassabis¹, Jue Wang¹, Sven Gowal¹ & Pushmeet Kohli¹ 

Abstract

Generative artificial intelligence (AI) models are revolutionizing biology, with tools such as AlphaFold 3 and protein design models accelerating breakthroughs in protein structure prediction and the creation of new functional proteins¹. Tracking and establishing the provenance of AI-generated protein sequences and structures is becoming increasingly important to tackle a range of emerging challenges, including biosecurity and concerns about information veracity^2,3,4. Here we introduce SynthIDBio, a family of methods for watermarking protein sequences and structures to establish the provenance of those generated with AI. SynthIDBio-sequence actively embeds a watermark into protein sequences while preserving function. We demonstrate this by creating watermarked, functional designed protein binders with binding affinity comparable with non-watermarked counterparts and near-perfect watermark detection accuracy. Furthermore, SynthIDBio-structure, a fine-tuned AlphaFold3 model, embeds an imperceptible watermark into biomolecular structures. Our work is a proof-of-concept that function-preserving biological watermarking is feasible, introducing a potential tool for provenance in the rapidly expanding era of AI-driven biological engineering.