특허 포트폴리오 분석을 통한 생성형 AI 기술의 투자 우선순위 도출 방법론 연구

Abstract

This study seeks to identify strategic investment priorities in generative AI field by categorizing key subcategories and analyzing patent positioning of leading countries. Leveraging a large language model (LLM), generative AI patents were classified into nine distinct categories. For each country, technology importance and technology attractiveness indices were computed, and the results were visualized using Ernst's patent portfolio analysis framework. Countries were then positioned into “Invest,” “Selective Investment,” and “Disinvestment” quadrants and corresponding investment strategies were proposed. The analysis revealed that while the United States maintains a leading position across most technology domains, countries such as Korea and China are positioned in quadrants where focused investments are required. This study offers a systematic framework for guiding investment decisions in generative AI technologies and highlights the feasibility of LLM-based patent classification as a tool for strategic patent management.