WeAreDevelopers Logo WeAreDevelopers
Sign in
For You LIVE
World Congress 2026 Europe • Jul 10, 2026 • Session details

Robust AIGC Provenance with Invisible Watermarking

Haoqi Wu

Standard metadata disappears during content sharing. Misinformation thrives in this gap. Discover how TikTok uses invisible, ML-based watermarks to embed indestructible AI provenance directly into pixel layers.

Pause
Mute Enter Fullscreen
#1 about 3 min

Introduction to AIGC provenance and authenticity challenges

The explosive growth of AI-generated media makes distinguishing synthetic content from reality increasingly difficult for users.

#2 about 3 min

Establishing proactive provenance and platform safety responsibilities

Recognizing its own generative output enables a platform to swiftly manage misinformation and ensure trusted safety.

#3 about 3 min

Implementing C2PA metadata for generative AI transparency

Attaching cryptographic manifests and visible labels to native generations offers a publicly verifiable lineage standard.

#4 about 2 min

Why cryptographic metadata degrades during media distribution

File-level credentials often fail to survive across platform re-encodings, aggressive compression, and user-initiated edits.

#5 about 2 min

Complementing C2PA metadata with robust invisible watermarking

Embedding a resilient, machine-readable signal directly into content frames prevents data loss when metadata is stripped.

#6 about 2 min

End-to-end workflow for injecting a recoverable UUID

Generating, distributing, and ultimately extracting an embedded UUID establishes a dependable recovery path for the original provenance context.

#7 about 4 min

Evaluating techniques for embedding robust media signals

Compared to traditional frequency domain designs, a machine learning strategy best resists compounding real-world adversarial transformations.

#8 about 4 min

Architecture of machine learning-based watermark neural networks

Utilizing joint encoder-decoder training within a shared latent space conceals provenance data as an unnoticeable pixel residue.

#9 about 2 min

Balancing robustness, imperceptibility, and bit capacity constraints

Delivering a scalable production watermark requires navigating the inherent tension between signal durability, data payload, and visual quality.

#10 about 2 min

Achieving visual imperceptibility with specialized loss functions

Employing pixel, regional, and temporal loss equations minimizes visible artifacts without compromising the underlying embedded identifier.

#11 about 2 min

Training against media augmentations to preserve signal integrity

Exposing the neural network to diverse cropping, compression, and filtering permutations ensures the signal outlasts hostile processing.

#12 about 1 min

Closing the automated provenance loop in production systems

Integrating dual embedding logic transforms a fresh upload pipeline into a reliable automated detector of recycled synthetics.

#13 about 2 min

Future steps for cross-platform open authenticity standards

Expanding user-facing verification tools and fostering industry interoperability will sustain media authenticity across varied internet ecosystems.

#14 about 2 min

Defending against adversarial attacks and dynamic circumvention tactics

Continually updating embedding policies and extraction model strategies outpaces malicious attempts designed to destroy the digital fingerprint.

Matching moments

53 sec

Tracing content provenance with the C2PA standard

Jen Looper Jen Looper · Europe 2026 Virtual

3:58 min

Protecting intellectual property and content verification with blockchain

Brian Whippo Brian Whippo · Coffee With Developers

1:44 min

Addressing the authenticity of AI translated video content

Chris Heilmann Chris Heilmann +2 · LIVE

8:22 min

Regulating deepfake content and AI transparency in Europe

Chris Heilmann Chris Heilmann +2 · LIVE

2:52 min

Audience questions on deepfake detection, regulations, and ethics

Cheuk Ho · World Congress 2023

8:37 min

AI video generation tools and content creation challenges

Chris Heilmann Chris Heilmann +2 · LIVE

More from World Congress 2026 Europe

Building the Nervous System of AI - Michael Kagan (NVIDIA) thumbnail
29:17
3 months ago

Building the Nervous System of AI - Michael Kagan (NVIDIA)

Michael Kagan, Sead Ahmetović

Five Things in Tech that Matter Now - WeAreDevelopers World Congress 2026 Closing Keynote thumbnail
19:41
3 months ago

Five Things in Tech that Matter Now - WeAreDevelopers World Congress 2026 Closing Keynote

Christian Heilmann

The Sound of Your Secrets: Teaching Your Model to Spy, So You Can Learn to Defend thumbnail
32:06
3 months ago

The Sound of Your Secrets: Teaching Your Model to Spy, So You Can Learn to Defend

David vonThenen

Super scaling for the Super Bowl: How to survive 30 million users hitting your backend in 30 minutes thumbnail
35:13
3 months ago

Super scaling for the Super Bowl: How to survive 30 million users hitting your backend in 30 minutes

Irina Branovic

Running Secure Life Science Research at Scale using Hybrid GPU HPC and Kubernetes 🧬 thumbnail
24:23
3 months ago

Running Secure Life Science Research at Scale using Hybrid GPU HPC and Kubernetes 🧬

Jeremy Murray

Angstfreude - AI and Corporate Culture - The Thrill and the Threat thumbnail
30:26
3 months ago

Angstfreude - AI and Corporate Culture - The Thrill and the Threat

Axel Ebert, Karin Krobath

Related videos

AI'll Be Back: Generative AI in Image, Video, and Audio Production thumbnail
29:54
World Congress 2025 • about 1 year ago

AI'll Be Back: Generative AI in Image, Video, and Audio Production

Anonymized By Request, Martin Foertsch, Thomas Endres

Overcome your trust issues! In a world of fake data, Data Provenance FTW thumbnail
26:41
World Congress 2023 • almost 3 years ago

Overcome your trust issues! In a world of fake data, Data Provenance FTW

Jon Geater

TikTok's Privacy Innovation thumbnail
20:46
World Congress 2024 • about 2 years ago

TikTok's Privacy Innovation

Mingshen Sun

The shadows that follow the AI generative models thumbnail
28:30
World Congress 2023 • about 3 years ago

The shadows that follow the AI generative models

Cheuk Ho

WeAreDevelopers LIVE - Building The World’s Worst Image Editor™ thumbnail
58:02
LIVE • 6 months ago

WeAreDevelopers LIVE - Building The World’s Worst Image Editor™

Chris Heilmann, Daniel Cranney, David Benson

GenAI Security: Navigating the Unseen Iceberg thumbnail
24:33
World Congress 2025 • 11 months ago

GenAI Security: Navigating the Unseen Iceberg

Maish Saidel-Keesing

Related articles

From learning to earning

Jobs that call for the skills explored in this talk.