The Korean Telecom Giant at the Center of Anthropic’s Mythos Controversy
A Korean telecommunications company, SK Telecom, found itself at the heart of a high-stakes geopolitical dispute in early 2025, as the Trump administration raised concerns over its access to Anthropic’s most advanced AI model, Claude Mythos. The incident unfolded amid broader anxieties about AI’s national security implications, revealing the tangled web of international business and technological influence that now defines the AI landscape. The controversy highlights the growing complexity of managing AI technologies that are both strategically vital and potentially vulnerable to misuse.
The Export Controls That Sparked a Controversy
The controversy began when Anthropic granted SK Telecom access to Claude Mythos, a powerful AI model capable of identifying software vulnerabilities and processing complex tasks. This decision, made as part of Anthropic’s Project Glasswing — a program designed to vet and expand access to its most advanced models — drew immediate scrutiny from the White House. US officials reportedly feared that SK Telecom’s ties to Chinese companies could compromise the security of the AI technology.
While SK Telecom itself has limited operations in China, its parent company, SK Group, has long maintained business interests in the country. These include partnerships with Chinese state-owned enterprises such as China Unicom, which were established more than two decades ago. The involvement of SK Telecom in these ventures, however minimal, became a focal point for US regulators seeking to control the flow of advanced AI technology.
A Tangled Web of Interests and Concerns
Anthropic’s decision to grant access to SK Telecom was not made lightly. The company had already engaged in extensive collaboration with the US government and cybersecurity experts to ensure that its models were not misused. SK Telecom, one of South Korea’s largest wireless carriers, had also invested $100 million in Anthropic in 2023 and was part of a commercial partnership aimed at developing AI solutions for the telecommunications industry.
Despite these precautions, the White House requested Anthropic to revoke access to Mythos for all foreign nationals, citing concerns over national security. This led to a broader decision by Anthropic to disable access to its most advanced models entirely, rather than attempt to enforce a nationality-based access policy.
The Aftermath and Ongoing Dispute
The fallout from the decision has left Anthropic and the Trump administration at odds. While Anthropic claims it is complying with the government’s requests, it has not fully resolved the underlying issues that led to the initial restrictions. The company also faces pressure from Amazon, which recently highlighted security vulnerabilities in a public version of the model, Fable 5.
As the debate over AI regulation continues to intensify, Anthropic remains caught between the demands of the US government and the realities of a global technology ecosystem. The dispute over SK Telecom and Claude Mythos is not just about access to AI, but about the broader question of how to govern technology that is increasingly powerful, opaque, and strategically significant.
Looking ahead, the situation underscores the growing complexity of AI governance. As the White House and Anthropic continue to negotiate, the industry faces a critical juncture. The incident highlights the urgent need for clear, internationally recognized standards to ensure that AI’s benefits are realized without compromising national security. Whether Anthropic can find a path forward that satisfies both its partners and regulators remains to be seen, but one thing is clear: the stakes are higher than ever.
