From deepfakes to DNA: the science of watermarking AI @googledeepmind
From deepfakes to DNA: the science of watermarking AI  @googledeepmind
Uploaded October 2026 | Updated November 2026, 3 hours ago
How do you know if what you're seeing is real? Professor Hannah Fry sits down with Pushmeet Kohli (VP of Science) and Jeremy Ratcliffe (Research Scientist) to break down SynthID, the watermarking technology expanding from text, images, and video straight into biology. Learn how invisible, mathematical watermarks help trace AI provenance, protect against misuse, and safeguard protein sequence design without impacting biological function.

Timecodes:
00:00 Introduction
00:34 What is a watermark?
04:35 SynthID
15:16 Images and video
19:28 SynthID Bio
28:00 Future of resilience

Learn more about SynthID Bio: https://deepmind.google/blog/introducing-synthid-bio/

___

Subscribe to our channel youtube.com/@googledeepmind
Find us on X https://x.com/GoogleDeepMind
Follow us on Instagram instagram.com/googledeepmind
Add us on Linkedin linkedin.com/company/deepmind
From deepfakes to DNA: the science of watermarking AIProject Genie | RollerballGemini Robotics: Dynamic interactionsVeo 3.1 - Designed to empower creativesGemini 3 Deep Think: Optimizing 2D semiconductor fabricationGemini 2.5 Pro Deep Think Demo | Competitive coding problemGemini 3 Deep Think: Optimizing 2D semiconductor fabricationShould you say ‘please’ when using AI?Gemini 3: Code a 3D visualization of the universeGemini Robotics Demo: Desk OrganizationUsing AI to manage resources in Africa | AI by you - Arnol’s storyVeo 3 demo | Magical origami
Google DeepMind |

From deepfakes to DNA: the science of watermarking AI

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER