As the use of Artificial Intelligence (AI) in design creation continues to expand, the Ministry of Intellectual Property (“MOIP”) recently released guidelines to clarify the standards for registering AI-assisted designs (the “Guidelines”). These Guidelines were prepared to mitigate confusion among industry applicants navigating rapid advancements in AI, and help them secure valid design rights.

The core principle of the Guidelines is that while AI-assisted designs are registrable, AI should only have been used as a tool in the design creation process. Under the current Design Protection Act, “the person who creates a design or his/her successor shall have the right to obtain a design registration”, which means that only a natural person (a human) can be recognized as the creator. Accordingly, the Guidelines stipulate that only individuals can be named on an application as the creator. If an AI system name - such as ChatGPT or Midjourney - or “AI developers” is identified as the creator, the application will be rejected. 

Furthermore, for an AI-assisted output to be a valid design right, the creator must have made a “substantial contribution” to the output. According to the Guidelines, results obtained by inputting broad, simple commands into an AI, such as “Make a cool chair design,” will not be recognized as reflecting human creativity. Conversely, a substantial contribution will be acknowledged if an individual used specific prompts that directly affect the shape, pattern, or color of the design, or creatively refined the AI-generated output to aesthetically enhance it.

Under existing practice, if an examiner has a reasonable basis to conclude that the applied-for design was created solely by AI or without substantial human contribution, the examiner may issue a preliminary rejection and request information from the applicant about the creation process. Since legal disputes could also potentially arise, the Guidelines recommend that applicants carefully document and maintain evidence that prove there was human intervention during the development of the design. Comprehensive records such as initial idea sketches, details of the AI models used, logs of prompt inputs and modifications, and the subsequent retouching process applied to the AI-generated outputs should be retained.

The Guidelines also highlight several invalidation risks that arise through the use of AI to create designs. These include the risk of a design being publicly disclosed and novelty being lost during the AI’s learning (training) process; the risk of the AI generating a design that is overly similar to an existing prior art; and the potential for structural inconsistencies between the multi-view drawings (e.g., front, side, and rear views) generated by the AI.

The Guidelines do not, however, introduce a duty to disclose to the MOIP whether the application relates to an AI-assisted design and practically speaking, the likelihood of a preliminary rejection being issued by an examiner is low since identifying AI intervention in the creation process is still challenging. However, there is a latent risk of a dispute, such as an invalidation action arising after registration. The design right could be invalidated if the creation process cannot be proven through legal discovery or if the lack of human intervention is exposed by an internal whistleblower. Therefore, for long-term, stable intellectual property protection, applicants should ensure their internal AI usage policies align with the intent of the Guidelines and systematically monitor and document their design creation processes.