Last Updated: 2026-05-29
UnMark uses artificial intelligence (AI) and machine learning (ML) technology to provide video watermark removal and content processing services. This AI Disclosure Statement describes our AI system's capabilities, limitations, and our commitment to transparency in accordance with applicable laws, regulations, and best practices, including guidance from the U.S. Federal Trade Commission (FTC) on AI marketing claims. This statement supplements Section 6 (Intellectual Property and DMCA) of our Terms of Service.
UnMark's core functionality is powered by proprietary deep learning models trained on extensive video data to automatically detect, isolate, and remove watermarks, subtitles, logos, timestamps, and other overlaid elements from video content. Our AI models employ computer vision techniques including object detection, inpainting, and frame reconstruction to preserve original video quality. Our AI models undergo continuous research, iteration, and optimization to improve processing accuracy, speed, and robustness across diverse video types and formats.
Our AI models achieve high accuracy rates on standard watermark types in common video formats. However, we do not guarantee 100% perfect results in all scenarios. AI processing output may exhibit the following known limitations: (a) Reduced accuracy on heavily compressed, low-resolution, or noisy source videos; (b) Residual artifacts when processing semi-transparent, animated, or complex multi-layered watermarks; (c) Performance variations with uncommon or proprietary video codecs and formats; (d) Potential degradation in output quality when the watermark occupies a significant portion of the frame. We strongly recommend reviewing all AI-processed outputs before use. While we continuously work to improve our models, these limitations are inherent to current AI technology and should be expected.
AI processing results provided by UnMark are generated by automated systems for general informational and creative purposes only. They do not constitute, and should not be relied upon as, professional advice of any kind (including but not limited to legal, medical, financial, or technical advice). You assume full responsibility for evaluating the accuracy, completeness, and suitability of AI-generated outputs for your intended use case. UnMark disclaims all liability for decisions made or actions taken based on AI processing results.
Your uploaded video content, metadata, and processing results are NOT used to train, fine-tune, improve, or refine our AI watermark removal models under any circumstances. Our AI models are trained exclusively on proprietary datasets, licensed content, and synthetic data that we generate internally. We maintain strict data isolation between user processing requests and our model training infrastructure. Your video content is used solely to fulfill your specific watermark removal request and is automatically deleted from our processing servers within 24 hours after processing completion. No human reviews your video content as part of our normal processing workflow. For complete details, see Section 1.3 (Video Content and Processing Data) of our Privacy Policy.
We are committed to developing AI systems that perform consistently and fairly across diverse user populations and content types. However, as with all machine learning systems, our AI models may exhibit performance variations attributable to the characteristics and distribution of training data. We actively monitor for systematic bias in model outputs and maintain a process for investigating and addressing reported fairness concerns. If you observe patterns of inconsistent or unfair AI behavior, please report them through the support channels listed in Section 7 below.
In accordance with FTC guidance on the use of AI in consumer products and services, UnMark makes the following transparency commitments: (a) We do not make unsubstantiated or misleading claims about our AI capabilities — known limitations are disclosed in Section 2 above; (b) We do not represent our AI as possessing human-like understanding, reasoning, or consciousness; (c) Our marketing materials and public communications about AI features are designed to be accurate, substantiated, and not deceptive; (d) We do not use AI to generate synthetic media or deepfakes — our AI performs content removal (subtractive processing) only, not content generation; (e) We comply with the DMCA and respect intellectual property rights — using our Service to remove watermarks from copyrighted content without authorization may constitute copyright infringement as detailed in our DMCA Copyright Policy and Section 13 (DMCA Copyright Policy) of our Terms of Service.
If you have questions, concerns, or feedback about our AI features, including reports of accuracy issues or potential bias, please contact us at support@mail.unmarkvideo.com or visit our Help Center at https://unmarkvideo.com/support.
