Abstract

As the DeepSeek model, which was released on January 20, 2025, has served as a starting gun for the paradigm shift in the AI and platform hegemony competition and the acceleration of ecosystem development from “big and expensive AI” to “smart and economical AI,” we have reviewed the strategies of major companies and the direction of platform policies.
In competition policy, the focus is currently on the upstream of the value chain (model development), where there is a high risk of competition, but it is necessary to pay attention to the downstream (AI applications) market, which is expected to develop more rapidly in the future. This is important to promote the development of open AI and respect for rights, to give small and medium-sized businesses and startups the opportunity to succeed, and to enable companies and users to access various innovative models.
In user protection policy, as competition for services using AI is in full swing, it is necessary to strengthen the legal and ethical foundations for improving service reliability and responding to side effects in a society-wide and systematic manner. Currently, the focus is on user deception such as AI washing, but there is a need to respond to the risks of general-purpose AI, such as the soon-to-be-spread AI agent.
As measures to strengthen the competitiveness of domestic AI and platforms, it is necessary to expand and accelerate infrastructure support for model development to secure cost-effective AI development capabilities, support efficient algorithm research and development, support domain-specific joint AI development, and To promote the convergence and utilization of data, the government proposed the establishment of a data-based system that responds to cost-effective AI, the establishment of a data cooperation network for major industries, and the creation of a generative AI sandbox and fostering of AI customization companies to support the rapid introduction and commercialization of AI.