딥러닝에서 버티칼 AI까지 - 지식재산정보서비스업에서 AI 솔루션 개발의 기회와 위험 요인 - = From Deep Learning to Vertical AI - Opportunities and Risks of Developing AI Solutions in the Intellectual Property Information Service -

다국어 초록 (Multilingual Abstract)

The intellectual property information service is one of the industries where AI utilization is a key source of innovation and has the greatest potential to lead to the discovery of new growth engines.
The purpose of this study is to examine the opportunities and risks of developing AI solutions in the intellectual property information service. To this end, this study divided the environmental changes resulting from the advancement of AI technology into three periods (development of deep learning and verification of AI performance, advancement of natural language processing(NLP) technology and expansion of Large Lanuage Model(LLM), and commercialization of generative AI). We analyzed opportunities for AI solution development, execution of those opportunities, and the risks that had to be taken in the process. As a result, in response to the advancement of AI technology and environmental changes, intellectual property information service companies have been gradually expanding their AI-based intellectual property information service solutions, from the development and renewal of patent classification and trademark similarity judgment solutions using deep learning technology to generative AI platform services and vertical AI, and have achieved AI transformation.
Furthermore, it was found that intellectual property information service companies have been overcoming the lack of internal and external resources required for AI solution development, the challenge of balancing the increasing costs of AI infrastructure expansion with investments to attract customers, and the threat of leaks of sensitive information and confidential information. To further expand economic performance through the development and utilization of vertical AI solutions in the intellectual property information service, government support that reflects the characteristics of the industry is necessary, along with independent efforts and investments by intellectual property information service companies to ensure AI reliability and safety.