Science
Method to ‘watermark’ AI-designed proteins could deter bioweapons, protect scientific credit
Publication: Latest News from Science Magazine
Author: Catherine Offord
Source published: 30/09/2026, 11:00:00 UTC
Briefing created: 30/09/2026, 16:05:17 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 30/09/2026, 15:57:25 UTC.
Briefings
A Google DeepMind team developed a method to embed a unique, identifiable pattern into the amino acid sequence of AI-designed proteins. This technique allows researchers to trace the origin of these proteins, which could deter bioweapons by revealing their creators, while simultaneously protecting the scientific credit of the original researchers who designed them. The study involved inserting a specific sequence into a protein without altering its structure or function, enabling identification without compromising safety.
Key points
- Google DeepMind successfully inserted an identifiable pattern into the amino acid sequence of AI-designed proteins.
- The method preserves the protein's structure and function while allowing for the identification of its creator.
Context and limitations
The study covers only AI-designed proteins; it does not address whether the method applies to naturally occurring proteins or biological systems.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 30/09/2026, 16:05:17 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeid2qnajngcalryvquadembgftmpep77ddih6tukxirr4sd2b3gtbe