Welcome to this brief AI watermarking chronology. Enjoy!
AI Watermarking Development Pre-2020
- 1999: Petitcolas et al. publish early work on information hiding techniques
- 2001-2003: Atallah et al. introduce early natural language watermarking using parsed syntactic tree structures
- 2006: Topkara et al. develop ambiguity-based watermarking through synonym substitutions
- 2015: Haribabu et al. propose the first robust image watermarking model using autoencoders
- 2017: Uchida et al. introduce the first method for embedding watermarks into deep neural networks
- 2018: Zhu et al. develop HiDDeN, achieving generation of visually indistinguishable watermarked images
- 2018: Zhang et al. propose methods for protecting intellectual property of deep neural networks with watermarking
- 2018: Adi et al. introduce backdoor-based watermarking for neural networks
- 2019: Rouhani et al. develop DeepSigns, an end-to-end watermarking framework for ownership protection
AI Watermarking Development 2020-2022
- 2020: Tancik et al. introduce StegaStamp for embedding bit-string watermarks in photos
- 2021: Jia et al. propose Entangled Watermark Embedding (EWE) for defending against model extraction
- 2021: Szyller et al. introduce DAWN, a dynamic adversarial watermarking system for neural networks
- 2022: Chakraborty et al. present DynaMarks, a system for defending against deep learning model extraction
AI Watermarking Development 2023
- 2023: Kirchenbauer et al. introduce “A Watermark for Large Language Models,” a statistical approach for text watermarking
- 2023: Zhao et al. propose “A Recipe for Watermarking Diffusion Models”
- 2023: Fernandez et al. introduce Stable Signature for watermarking latent diffusion models
- 2023: Luo et al. develop CopyRNeRF for protecting copyright of Neural Radiance Fields
- 2023: Wen et al. propose Tree-Ring watermarks as fingerprints for diffusion images
AI Watermarking Development 2024
- 2024: Feng et al. introduce AquaLoRA for watermarking Stable Diffusion models
- 2024: Ci et al. develop WMAdapter for adding watermark control to latent diffusion models
- 2024: Ci et al. extend Tree-Ring to RingID for multi-key identification
- 2024: Jang et al. introduce WateRF for robust watermarking in radiance fields
- 2024: Huang et al. develop GaussianMarker for copyright protection of 3D gaussian splatting
- 2024: Zhang et al. create GS-Hider for hiding messages into 3D gaussian splatting
- 2024: Li et al. introduce GaussianStego for generative 3D gaussians splatting
- 2024: Yang et al. develop Gaussian Shading for performance-lossless image watermarking
- 2024: Lei et al. introduce DiffuseTrace, a flexible watermarking scheme for latent diffusion models
- 2024: Jiang et al. develop SmartMark for watermarking smart contracts on blockchain platforms
- 2024: Dziembowski et al. introduce VIMz for private proofs of image manipulation
AI Watermarking Development 2025
- 2025: Xu et al. develop robust multi-bit text watermarking with LLM-based paraphrasers
- 2025: Chao et al. introduce the Robust Binary Code (RBC) watermark using error-correcting codes
- 2025: Kulthe et al. create MultiNeRF for embedding multiple watermarks in Neural Radiance Fields
- 2025: Li et al. introduce GaussianSeal for 3D Gaussian Generation Model watermarking
- 2025: Petrov et al. discover watermark coexistence and develop watermark ensembling techniques
- 2025: Bagad et al. present zkDL++, a framework for cryptographic watermark verification using zero-knowledge proofs
- 2025: Sharma and Kim develop frameworks combining digital watermarking with blockchain and perceptual hash functions
- 2025: Feng et al. introduce integrated approaches combining blockchain and watermarking technologies
- 2025: Luo et al. publish a comprehensive survey on digital watermarking technology for AI-generated images
Thanks for reading!