AI를 이용하여 작성한 특허명세서의 실효성 확보 방안

• 영문명: Method to Ensure the Effectiveness of AI-Generated Patent Applications
• 발행기관: 충북대학교 법학연구소
• 저자명: 함영욱(Young Wuk HAM)
• 간행물 정보: 『법학연구』第36卷 第1號, 415~445쪽, 전체 31쪽
• 주제분류: 법학 > 법학
• 파일형태: PDF
• 발행일자: 2025.06.30

국문 초록

Generative artificial intelligence or large language models can create descriptions of inventions and claims sections of patent applications by receiving data related to the content of the invention. These generated descriptions of inventions and claims sections, which are included in the specifications, drawings, or summaries attached to patent applications, are referred to as “AI-written patent specifications.” Major patent offices in countries such as Europe, the United States, China, Japan, and South Korea permit the submission of these AI-written patent specifications. There may be issues related to technical leaks concerning the data input into generative artificial intelligence or large language models for the purpose of drafting patent specifications related to the content of the invention. Additionally, there are limitations associated with generative artificial intelligence or large language models, such as the inclusion of hallucinations in the AI-generated patent specifications, the limitation of the scope of rights in the AI-generated patent specifications, and the possibility that the claims included in the AI-generated patent specifications may differ from the content of the data related to the invention provided by the applicant or patent attorney. To address these issues, it is necessary to introduce a regulation mandating the indication of AI authorship in Article 42, Paragraph 1, Item 1, implement a requirement for the submission of the applicant's consent in Article 7, Paragraph 2, and establish a new Article 7-3 to introduce a mandatory review by a patent attorney.

목차

Ⅰ. 서론

Ⅱ. AI를 이용하여 작성한 특허명세서의 인정 여부

Ⅲ. AI를 이용하여 작성한 특허명세서 인정 시의 문제점

Ⅳ. AI를 이용하여 작성한 특허명세서의 인정 방안

Ⅴ. 결론

참고문헌

키워드

특허법 인공지능기본법 발명 주체 인공지능 권리범위 환각 Patent Act AI ACT Inventor Artificial intelligence generative ai Large Language Model patent application Hallucination

참고문헌

• 법문사
• KOSEN REPORT 2023
• 산업융합연구
• 성균관법학
• 공법연구
• 학습자중심교과교육연구
• The Journal of Law&IP
• 법학논문집