[Anchor]
Anthropic, a major U.S. AI company and key competitor to OpenAI, has opened an office in Seoul and held a press conference.
Attention was particularly focused on recent export control measures by the U.S. government. Reporter Kwak Sang-eun was there.
[Reporter]
At the press conference announcing the entry of U.S. AI firm Anthropic into the Korean market, questions poured in regarding the controversy over U.S. government export controls on Anthropic's latest conversational AI models, Mythos 5 and Fable 5.
Concerns have been raised that these models could be misused for hacking, as they are capable of identifying software vulnerabilities.
For this reason, Anthropic had granted early access to select trusted companies and institutions for the purpose of international cooperation in building cybersecurity systems. However, on June 12, the U.S. administration completely banned access to Mythos 5 and Fable 5 by foreign institutions and individuals, citing "national security needs."
On June 15, The Washington Post reported that this export control measure was prompted by suspicions that a Korean company with ties to China had gained access to the models.
Anthropic has currently suspended the service itself.
An executive in charge of Anthropic International expressed hope that the export control issue would be resolved soon, but remained tight-lipped about specific details.
[Chris Chauri / Head of Anthropic International: I think this will be sorted out in a few days. I don't think the export controls will remain in place.]
Anthropic stated that the Korean AI market, which has a robust government regulatory framework, aligns well with its "safety-first" philosophy, and expressed high expectations for the rapid growth of the Korean market.
Furthermore, the company announced that through the opening of this Seoul office, it would more effectively respond to the needs of domestic corporate clients and strengthen support for academia and the public sector.
(Video reporting: Park Jin-ho)
※ Please note: This article was translated by AI and may contain errors.