▲ Anthropic
The U.S. Trump administration once considered export controls on Anthropic after the company allowed certain "China-linked entities" early access to its latest artificial intelligence (AI) models, "Mythos 5" and "Fable 5," before their official release, The Washington Post (WP) reported on June 15 (local time).
According to White House officials cited by the WP, Anthropic submitted a list of 111 organizations granted early access to its latest AI model, "Mythos," to the Trump administration several weeks ago, which the administration reviewed and approved.
However, Anthropic later disclosed that approximately 50 additional organizations had been granted access. When the company did not reveal the identities of the newly added organizations for several days, the administration began considering export controls.
The Trump administration eventually discovered that the list provided by Anthropic included a South Korean telecommunications company suspected of having ties to China, officials said.
Prior to the official launch of "Mythos 5" and "Fable 5," Anthropic had allowed certain domestic and international institutions and companies to use Mythos early through "Project Glasswing."
Project Glasswing is a global cybersecurity coalition launched by Anthropic when it announced the development of Mythos.
It has been operated by providing the model in advance to verified companies and institutions to defend against software vulnerabilities, amid concerns that Mythos could be misused by malicious hackers.
A source familiar with the matter told the WP that Anthropic promptly revoked the South Korean company's access to Mythos.
However, the incident reportedly undermined the confidence of Trump administration officials in Anthropic's ability to protect sensitive technology.
The WP reported, "The Trump administration had already been considering export controls for weeks before forcing the company to suspend services, following the controversy over sharing technology with a company suspected of having ties to China."
On June 12, the U.S. government announced export control guidelines prohibiting foreign institutions and individuals from accessing Anthropic's top-tier AI models, "Mythos 5" and "Fable 5."
In compliance with these guidelines, Anthropic immediately suspended "Mythos 5" and "Fable 5" services for all users.
(Photo: Yonhap News)
※ Please note: This article was translated by AI and may contain errors.