AI 대전환의 시대: 거대 모델의 개방과 기업의 생존 전략 > K-wave Trends

The emergence of Antropic's newly introduced top-level model 'Mythos 5' and general-purpose model 'Fable 5' presents a sophisticated balance between AI safety and performance. Antropic has dualized its model to provide overwhelming intelligence in general business areas while introducing conservative safeguards in sensitive areas such as cybersecurity and biohazards. In particular, the Mythos 5 model, which is provided only to organizations with verified security, is expected to serve as a bridge for major domestic companies such as Samsung Electronics and SK Hynix to utilize cutting-edge AI technology in a safe environment. Strengthened data retention policies and dramatic improvements in performance indicators suggest that an environment is being created where companies no longer have to hesitate to introduce AI due to hallucinations or security concerns.

Samsung Group's declaration of an 'AI transformation' comparable to the first digital transformation in 30 years has significant implications for the domestic industry. Samsung, which had previously restricted the use of externally generated AI due to concerns about information leaks, is now moving toward open innovation by officially introducing models from global big tech companies such as ChatGPT, Gemini, and Claude throughout the company. Intensive training programs such as 'AX Boot Camp' for the entire CEO team go beyond simply borrowing tools and demonstrate a strong will to increase management's AI literacy and reorganize the entire organization's DNA into an AI-centered system. This is a concrete step to innovate the entire value chain, including manufacturing, marketing, and management support, based on AI, and will be a new milestone in the way large domestic companies adopt AI.

With the spread of AI technology, the global regulatory environment surrounding it and competition to build infrastructure are becoming more intense. The European Union's order to allow Meta's WhatsApp to allow access to rival AI assistants shows regulators' resolute determination to check monopolistic AI services that leverage their platform dominance. Meanwhile, OpenAI is making every effort to secure hardware infrastructure, including collaborating with NVIDIA to secure an ultra-large data center worth 762 trillion won. This battle over capital power and infrastructure is placing a huge financial burden on AI developers, and this is expected to become the most important variable in the IPO process of AI companies or in evaluating corporate soundness in the future.

The Korean government and companies are accelerating the construction of their own ‘physical AI’ ecosystem to respond to this global trend. A consortium led by LG Electronics and KT is developing a 'world model' that learns physical laws in the virtual world and is taking a strategy of internalizing massive data from manufacturing sites into AI. This aims to improve practical productivity by going beyond text-oriented AI and controlling actual robots and industrial facilities. In particular, the plan to release developed technologies as open source so that even small and medium-sized businesses can benefit is expected to greatly contribute to expanding the base of the domestic AI ecosystem and reducing dependence on foreign technologies.