Mint Corporation AI Law Response & Invisible Watermarking

In line with the implementation of the ‘AI Basic Act’, the Korea Mint and Security Council has applied core technologies that strengthen the transparency of AI products and digital content copyright protection to private services. The company’s plan is to expand the security technology it has accumulated into the AI environment and build a trust base for the rapidly increasing number of generated contents.

The Korea Mint and Mint Research Institute announced that it had applied its self-developed ‘AI-based invisible digital watermark’ technology to ‘LOOK PICK AI’, an AI virtual fitting and lookbook creation service. It is evaluated as an example of the public sector responding to the deepfake and unauthorized theft problems that are deepening with the spread of AI, and proactively responding to institutional changes requiring identification of AI products.

Invisible watermarking is a method of inserting a security code that is not visible to the human eye distributed throughout the image. Dedicated verification tools allow you to immediately determine whether an image was created with AI while maintaining visual quality. Compared to visible watermarks, it has high security in that it is difficult to remove and is resistant to forgery and falsification. In particular, identification information is maintained not only in digital images but also in printed materials, so it has the potential to expand into the realm of physical products such as packages and labels.

Lookpick AI, to which this technology is applied, is a commercial platform that creates AI model wearing images based on clothing photos uploaded by users. Equipped with watermark technology, an invisible ‘digital authentication of authenticity’ is granted to service results, and users can prove that the image was created by a legitimate AI service and clearly claim ownership against unauthorized capture or commercial theft by a third party. It goes beyond the simple display of AI creation and combines the function of protecting content as a digital asset.

This application is significant in that it technically implements the core purpose of the AI Framework Act, ‘user recognition potential’. As damage to trust and market turbulence are inevitable in an environment where the creator and source cannot be identified, invisible watermarks are evaluated as an effective means of simultaneously supporting regulatory compliance and revitalizing the industry. In particular, it is highly likely to serve as a foundational technology for preventing copyright disputes in fields where AI images are highly utilized, such as fashion, webtoons, and entertainment.

Using this case as a starting point, the Korea Mint and Security Research Institute plans to expand technical support across various private platforms and industries to contribute to the creation of a trustworthy AI ecosystem. The strategy is to establish a safe and sustainable digital order even amid institutional changes by transplanting the public trust accumulated through currency and security technology to the AI content area.

*Below is the English article translated with ‘Google Translate’ [전문]no see. ‘Google Translate’ is working to improve understanding. It is assumed that there may be errors in the English translation.<*The following is [the full text] of the English article translated by 'Google Translate'. 'Google Translate' is working hard to improve understanding. It is assumed that there may be errors in the English translation.>

Korea Minting, Security Printing, and Printing Corporation (KOMSCO) Begins Full-Scale Response to AI Framework Act… Invisible Watermarking Implemented in Private Services

Simultaneously Implementing AI Product Identification and Copyright Protection… LookPick AI Incorporates ‘Digital Authenticity Certification’ Technology

Korea Minting, Security Printing, and Printing Corporation (KOMSCO) has applied core technologies to its private services to enhance the transparency of AI products and copyright protection for digital content, in line with the implementation of the “AI Framework Act.” This move aims to expand the corporation’s accumulated security technologies into the AI environment and establish a trustworthy foundation for the rapidly growing generational content.

KOMSCO’s Technology Research Institute announced the application of its proprietary “AI-based invisible digital watermarking” technology to “LOOK PICK AI,” an AI virtual fitting and lookbook generation service. This is seen as a proactive response by the public sector to address the growing issues of deepfakes and unauthorized theft with the proliferation of AI, and to regulatory changes requiring the identification of AI products.

Invisible watermarking involves distributing security codes invisible to the human eye throughout an image. While maintaining visual quality, dedicated verification tools can instantly determine whether an image was generated by AI. Compared to visible watermarks, this technology is more difficult to remove and more resistant to forgery and tampering, offering enhanced security. Furthermore, since identification information is maintained not only on digital images but also on printed materials, it holds promise for expansion into physical product areas such as packaging and labels.

LookPick AI, where this technology is applied, is a commercial platform that generates images of AI models wearing clothing based on user-uploaded photos. The watermark technology provides an invisible “digital authentication” to the resulting service output, allowing users to prove that the image was generated by a legitimate AI service and clearly assert ownership against unauthorized capture or commercial use by third parties. Beyond simply indicating AI-generated content, this technology combines the functionality of protecting the content as a single digital asset.

This implementation is significant in that it technically implements the core principle of the Framework Act on AI: “user recognizability.” In an environment where creators and sources cannot be identified, loss of trust and market disruption are inevitable. Therefore, invisible watermarking is considered an effective means of simultaneously supporting regulatory compliance and industry revitalization. In particular, it has great potential to serve as a foundational technology for preventing copyright disputes in areas with high AI image utilization, such as fashion, webtoons, and entertainment.

Based on this case, the Korea Minting and Security Printing Corporation’s Technology Research Institute plans to expand technical support across various private platforms and industries, contributing to the creation of a trustworthy AI ecosystem. By transferring the public trust built through currency and security technologies to the AI content sector, the company aims to establish a safe and sustainable digital order even amidst institutional changes.

View original article:Break News Daejeon Sejong ChungcheongView original article:Break News Daejeon Sejong Chungcheong