# DRIFT: Removing Diffusion Watermarks by Deflecting the Generative Trajectory

arXiv: https://arxiv.org/abs/2609.08213 · HTML: https://mlsanity.com/paper/2609.08213/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Watermarking generated content and models, one of 12 directions in LLM safety and security.
- Builds on 8 of its 42 references that are in the map.
- First posted on arXiv 8 Sep 2026.

## Placed in

- LLM safety and security / [Watermarking generated content and models](https://mlsanity.com/safety/watermarking-generated-content-and-models/)

## Builds on (most cited first, 8 of 8)

- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Robustness of AI-Image Detectors: Fundamental Limits and Practical Attacks](https://mlsanity.com/paper/2310.00076/) (2023)
- [An Undetectable Watermark for Generative Image Models](https://mlsanity.com/paper/2410.07369/) (2024)
- [Image Watermarks are Removable Using Controllable Regeneration from Clean Noise](https://mlsanity.com/paper/2410.05470/) (2024)
- [Schedule Your Edit: A Simple yet Effective Diffusion Noise Schedule for Image Editing](https://mlsanity.com/paper/2410.18756/) (2024)
- [SEEDS: Exponential SDE Solvers for Fast High-Quality Sampling from Diffusion Models](https://mlsanity.com/paper/2305.14267/) (2023)
- [Hidden in the Noise: Two-Stage Robust Watermarking for Images](https://mlsanity.com/paper/2412.04653/) (2024)
- [Forging and Removing Latent-Noise Diffusion Watermarks Using a Single Image](https://mlsanity.com/paper/2504.20111/) (2025)

## Abstract (arXiv)

Diffusion watermarking embeds verifiable signals into the generative process and commonly verifies them by recovering trajectory-dependent evidence, making the marks robust to conventional pixel-space distortions. Existing removal attacks either regenerate along deterministic trajectories, which often preserve the watermark-bearing latent structure, or optimize every image separately. We identify the reliance on a recoverable generative trajectory as a common attack surface among the schemes we study. Based on this observation, we propose DRIFT, a black-box attack that combines partial forward diffusion with stochastic reverse resampling. Forward re-noising limits source information available to a fixed-depth recovery pipeline, while stochastic reversal supplies alternative noise-driven paths whose removal benefit we isolate through matched sampler comparisons. Adaptive DRIFT searches a selected ladder for each image's first verifier-rejected rung and refines fidelity while retaining only updates rejected by the same verifier. At fixed depth, we derive information-theoretic and Wasserstein source-dependence bounds; under realized-ladder monotonicity, the first rejected rung is least distorted among rejected rungs on that ladder, and verifier-gated refinement preserves rejection. Across nine watermarks spanning three paradigms, DRIFT achieves 98-100% attack success and the best image quality among the compared attacks, without secret keys, verifier internals, or per-image gradient optimization.