U.S. artificial intelligence company Anthropic officially launched its Seoul office, marking its entry into the South Korean market. However, it did not provide specific answers regarding recent controversies surrounding access restrictions to its advanced AI models, Mythos 5 and Fable 5, as well as Project Glasswing.
On June 17, Anthropic held a press conference at the Conrad Hotel in Yeouido, where executives discussed the company's strategy for entering the South Korean market. The event featured Ankris Chauri, head of Anthropic International, and Choi Gi-young, CEO of Anthropic Korea. Tom Brown, co-founder and Chief Operating Officer, was unable to attend due to prior commitments in the U.S.
Chauri emphasized that South Korea's AI basic law aligns with Anthropic's philosophy of promoting safe technological development, expressing enthusiasm for the company's expansion into the country.
During the event, numerous questions arose regarding the access restrictions for Mythos 5 and Fable 5, as well as the status of Project Glasswing. Chauri stated, "The potential for bypassing safety measures or jailbreaking applies to all models in the industry. This is a limited case." He added, "Currently, there is nothing more we can share on this matter."
With the launch of its Seoul office, Anthropic aims to strengthen its connections with local businesses and the developer ecosystem. Prior to the press conference, the company held a private event inviting key representatives from major South Korean firms. Attendees included officials from Samsung Electronics, SK Hynix, and SK Telecom, among others in the AI and semiconductor sectors.
Choi noted, "We began the launch event for the Seoul office yesterday, discussing AI utilization and collaboration strategies with Korean companies." He added, "Tomorrow, we plan to hold a developer event to further engage with the local developer community."
According to Anthropic, South Korea ranks 12th globally in terms of per capita usage of Claude among 116 countries. Choi remarked, "This indicates a significant number of users are utilizing Claude, particularly in both technical and creative tasks."
Anthropic is also looking to strengthen partnerships not only with businesses and the developer community but also with research institutions and non-profit organizations. To support this, the company plans to provide Claude to researchers at the National AI Research Lab (NAIRL) to assist with key research in AI safety, model evaluation, alignment, and robustness. Choi stated, "We are researching safe AI development in collaboration with KAIST, Korea University, Yonsei University, and POSTECH, and we plan to provide Claude free of charge to researchers."
As a differentiation strategy from OpenAI, which has already entered the South Korean market, Choi highlighted the importance of partnerships with systems integrators (SI). He stated, "We will support developers in enhancing productivity and delivering results to customers more quickly through Claude, and strengthening SI partnerships will be a key competitive advantage."
Chauri reiterated, "Our focus on safety and the enterprise market is Anthropic's strength," asserting, "We are leaders in the enterprise AI market." He noted that the company's current market share in the enterprise sector is approximately 40%, with revenues reaching $47 billion during the last investment round.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.