This article presents the development of an image generation system that employs digital watermarking and metadata embedding technologies to determine whether an image has been generated by an AI model. The system acts as an intermediary service between providers (web services with generation models) and end users, ensuring seamless integration of these technologies. With the growing volume of AI-generated content, distinguishing such images from authentic ones has become increasingly challenging. Additionally, the lack of universal tools for managing generated assets and embedding metadata creates inefficiencies and risks related to authenticity and intellectual property. This article attempts to create a viable centralized solution that integrates protection measures into any user-generated image, regardless of the originating service. The system operates as a middleware solution compatible with existing generation models, providing a unified interface for users. Developed pipeline facilitates both addition of watermarking into the generative process as well as embedding metadata. The intuitive interface enhances usability, while the centralized repository enables users to manage and verify their generated content. This approach is innovative, combining digital watermarking, metadata integration, and centralized management within a single platform. Unlike existing tools tailored to specific platforms, this system offers cross-service functionality. The solution is highly relevant for content authenticity, intellectual property management, and user convenience. It enhances trust in digital content and provides a scalable architecture adaptable to diverse platforms and needs. Future research could extend this approach to broader areas of information technology, ranging from non-image generation models to operating system-level modules for protecting against generated products.
[1] Fernandez P, Couairon G, Jégou H, Douze M, Furon T. The stable signature: rooting watermarks in latent diffusion models. arXiv.org. https://arxiv.org/abs/2303.15435. Published March 27, 2023.
[2] Xu R, Hu M, Lei D, et al. InvisMark: Invisible and robust watermarking for AI-generated image provenance. arXiv.org. https://arxiv.org/abs/2411.07795. Published November 10, 2024.
[3] Jiang Z, Guo M, Hu Y, Gong NZ. Watermark-based attribution of AI-Generated content. arXiv.org. https://arxiv.org/abs/2404.04254. Published April 5, 2024.
[4] Fernandez P. Watermarking across Modalities for Content Tracing and Generative AI. arXiv.org. https://arxiv.org/abs/2502.05215. Published February 4, 2025.
[5] Li G, Chen Y, Zhang J, et al. Warfare:Breaking the watermark protection of AI-Generated content. arXiv.org. https://arxiv.org/abs/2310.07726. Published September 27, 2023.
[6] Padhi, S. K., Tiwari, A., & Ali, S. S. (2024). Deep learning-based dual watermarking for image copyright protection and authentication. IEEE Transactions on Artificial Intelligence, 1–12. https://doi.org/10.1109/tai.2024.3485519
[7] Fairoze, J., Ortiz-Jiménez, G., Vecerik, M., Jha, S., & Gowal, S. (2025, February 7). On the Difficulty of Constructing a Robust and Publicly-Detectable Watermark. arXiv.org. https://arxiv.org/abs/2502.04901
[8] Simmons, J. C., & Winograd, J. M. (2024, May 20). Interoperable Provenance Authentication of Broadcast Media using Open Standards-based Metadata, Watermarking and Cryptography. arXiv.org. https://arxiv.org/abs/2405.12336
[9] SynthID. (2025, February 25). Google DeepMind. https://deepmind.google/technologies/synthid/
[10] OpenAI joins C2PA Steering Committee - C2PA. (n.d.). https://c2pa.org/post/openai_pr/
[11] Balan K, Agarwal S, Jenni S, Parsons A, Gilbert A, Collomosse J. EKILA: Synthetic Media Provenance and Attribution for Generative Art. arXiv.org. https://arxiv.org/abs/2304.04639? Published April 10, 2023.
[12] Longpre S, Mahari R, Obeng-Marnu N, et al. Data Authenticity, Consent, & Provenance for AI are all broken: what will it take to fix them? arXiv.org. https://arxiv.org/abs/2404.12691v1? Published April 19, 2024.
[13] Bieniek J, Rahouti M, Verma DC. Generative AI in Multimodal User Interfaces: Trends, challenges, and Cross-Platform Adaptability. arXiv.org. https://arxiv.org/abs/2411.10234. Published November 15, 2024.
[14] Tolomei G, Campagnano C, Silvestri F, Trappolini G. Prompt-to-OS (P2OS): Revolutionizing Operating Systems and Human-Computer Interaction with Integrated AI Generative Models. arXiv.org. https://arxiv.org/abs/2310.04875. Published October 7, 2023.
[15] Luera R, Rossi RA, Siu A, et al. Survey of user interface design and Interaction Techniques in Generative AI Applications. arXiv.org. https://arxiv.org/abs/2410.22370v1. Published October 28, 2024.
[16] C2PA specifications :: C2PA specifications. https://c2pa.org/specifications/specifications/2.1/index.html.
[17] Bond-Taylor S, Leach A, Long Y, Willcocks CG. Deep Generative Modelling: a comparative review of VAEs, GANs, normalizing flows, Energy-Based and autoregressive models. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2021;44(11):7327-7347. https://doi.org/10.1109/tpami.2021.3116668