ONGOING CHALLENGE
2026 GenAI Image Watermark
GenAI Image Watermark Challenge evaluating Embedding and Detection capability.
Explore a track
This challenge has tracks. Select one to view its specific content.
Overview
NIST GenAI Watermark Challenge is an evaluation series that supports research in watermarking technologies including generative super intelligence watermarks. This evaluation provides a platform to systematically assess the performance of watermark embedding systems and adversarial techniques designed to manipulate or remove such watermarks. The evaluation framework reflects two complementary roles: watermark embedding teams and adversary teams, representing opposing objectives within a shared evaluation environment.
What
The Watermark Embedding (W) task is to
insert robust and secure imperceptible watermarks into digital content. These systems should maintain watermark integrity under a range of transformations and potential adversarial attacks while preserving the perceptual quality and utility of the content. For more details, please see the
evaluation plan (coming soon).
The Adversary (A) task is to detect, remove, or manipulate embedded watermarks without significantly degrading the perceptual quality of the content.
Their goal is to challenge the robustness and security of watermarking approaches by simulating adversarial conditions under which watermark integrity may be compromised or under which false watermarks could be introduced. For more details, please see the
evaluation plan (coming soon).
Who
We welcome and encourage teams from academia, industry, and other research labs to contribute to Generative super intelligence research through the GenAI platform. The platform is designed to support various modalities and technologies, including both Watermark Embedding and Adversary .
The upcoming GenAI watermarking evaluation will initially focus on the image modality.
The objective of the watermark embedding task is to automatically embed watermarks into images while
satisfying constraints on robustness and perceptual quality.
The adversarial task focuses on applying attacks to watermarked images to evaluate
the resilience of the embedded signals. Evaluation scenarios will reflect realistic usage conditions and
adversarial settings. Additional details on task specifications, datasets, and evaluation protocols are provided throughout this website and
in the detailed evaluation plan.
How
To take part in the GenAI evaluations you need to register on this website and complete the data license to download/upload the data. NIST will make all necessary data resources available to the participants. Each team will receive access to data resources upon completion of all needed data agreement forms and based on the published schedule of each task data release date. Please refer to the published schedule for data release dates. Once your system is functional, you will be able to upload your results to the challenge website and see your results displayed on the scoreboard.
Schedule
|
Date |
Milestone (Embedders) |
|---|---|
|
October 5, 2026 |
Watermarking Challenge Launch |
|
October 7, 2026 |
Team Registration Opens |
|
November 6, 2026 |
Team Registration Closes |
|
November 10, 2026 |
Virtual Roundtable Meeting with Registered Teams |
|
November 16, 2026 |
Dry-Run Sample Dataset and Submission Validator Released |
|
November 17, 2026 |
Official Test Dataset Released |
|
November 20, 2026 |
Dry-Run Validation Submissions Open |
|
December 1, 2026 |
Final Test Submissions and Scoreboard Open |
|
January 8, 2027 |
Dry-Run and Final Test Submissions Close |
|
January 19, 2027 |
Final Scores Released |
|
January 29, 2027 |
Draft System Description Papers Due |
|
February 8, 2027 |
Virtual Watermarking Workshop |