연합뉴스
| yna@yna.co.kr 2025-03-19 14:21:22
*Editor’s note: The number of global Hallyu (Korean Wave) fans is approaching approximately 225 million, according to the 2024 report by the Korea Foundation. The surge in fans marks the dawn of the "Digital Silk Road" era, where communication transcends the limitations of time and space, enabling real-time interaction across the globe. Truly, we are in the era of "Hallyu 4.0."
Suk Soo-sun's Design Management Story: AI and Intellectual Property Issues
Contributed by Suk Soo-sun (professor at Yonsei Graduate School of Communication & Arts)
◇ AI-Generated Images and Intellectual Property
The images generated by generative AI have recently become a significant issue for trademark owners. These AI systems are often built without considering intellectual property rights, leading experts to continuously raise concerns about this problem.
One representative "text-to-image" generation tool, DALL-E 2, modifies web images and generates data by learning the relationships between millions of texts and images. In addition to DALL-E 2, other similar AI tools such as Midjourney and Stable Diffusion are becoming more accessible, with all of them rapidly improving the quality of images.
Moreover, AI-based video generation technology is also receiving significant attention recently and is expected to enhance the cost-efficiency of marketing and content creation.
Despite these technological advances, regulation and management strategies have not kept pace with these changes. AI-generated images are becoming increasingly common on platforms like Shutterstock, iStock, and Adobe Stock. In particular, without clear standards for the copyright and ownership of images, brands may face significant confusion.
For instance, DALL-E 2 filters out pornography and duplicate images during its production process, yet it can still generate images that include iconic symbols or trademark characters.
AI can create images that include well-known characters like SpongeBob or Homer Simpson.
Recently, generative AI has been gaining attention in marketing. There is an ongoing discussion about the impact of AI on graphic design and content creation. AI-generated videos could become an affordable and efficient alternative to traditional advertising production.
Advertisers can create multiple ads with the same budget to target specific consumer groups.
However, AI-generated images present intellectual property and branding issues. For example, if an image generated by DALL-E 2 includes Disney characters or trademark logos, the brand is likely to claim copyright infringement.
In such cases, businesses need to establish legal and ethical standards while utilizing generative AI. It is necessary to discuss whether it could be legally problematic for AI systems to learn from copyrighted images. Furthermore, it is crucial to discuss how the spread of open-source AI platforms complicates this issue.
◇ The Advancement of AI Image Generation Technology and the Need for Legal Responses from Companies
With the rapid development of AI image generation technology, companies are increasingly required to strengthen their intellectual property (IP) protection and establish strategies based on new legal standards. These legal challenges could potentially hinder brand innovation, so prompt responses are urgently needed.
Companies must clearly define the boundary between AI-generated images and existing IP to prevent copyright and trademark infringements in advance. Such preparation will help minimize legal risks while maintaining the creativity and innovation of brands.
Efforts to maximize the opportunities of new technologies while preventing legal issues are essential. Ultimately, the development of AI suggests that brands must redefine and overcome the challenges they face through a forward-looking approach.
Brands and logos are at risk of being caught in voluntary AI experiments or the creation of humor and memes. These issues often arise not for profit but simply for fun, which contrasts with the experimental aspects of campaigns like Heinz's. Brands remain cultural icons, and the outcomes, whether good or bad, become a part of reality.
AI scientist Janelle Shane conducted an experiment on her blog AI Weirdness using DALL-E 2 to create disastrous versions of corporate logos. DALL-E 2 is an AI that generates images based on text prompts, using clips trained by OpenAI on various images and texts.
Through this experiment, DALL-E demonstrated excellent performance in generating clear and consistent images.
Shane began the experiment with simple requests, such as "local waffle house" or the Pizza Hut logo. Interestingly, the AI appears to have gathered enough information online to accurately reproduce the original logos. The result mixed elements of accidental satire with critiques of the relative strength of specific brands.
The creativity demonstrated by DALL-E provides an opportunity to explore the potential of AI in the design field. The advancement of AI technology brings about an intriguing process of reconstructing existing concepts and creating new visual interpretations.
Ultimately, AI tools like DALL-E 2 have the potential to expand the boundaries of design and contribute to the redefinition of the format and features of corporate logos.
Through such experiments, it is possible to explore whether AI’s creativity can transcend existing frameworks and open up new opportunities.
However, it remains uncertain whether a brand can make IP-based claims about these images. It is also difficult to be sure whether challenging such claims is a good idea.
So far, DALL-E and its competing tools have demonstrated an astonishing ability to visualize anything imaginable.
As a result, AI-generated images are realistic, and companies need to establish strategies that align with these changes. Defining the boundary between brand and creativity will be a key issue moving forward.
This is a critical time to carefully consider the opportunities and challenges that the advancement of AI will bring.
[ⓒ K-VIBE. 무단전재-재배포 금지]