Anthropic CEO meets White House officials amid cybersecurity model rollout

Tensions between Anthropic and the Trump administration may be easing following the AI company's introduction of a powerful cybersecurity model and a White House meeting with its CEO. Anthropic CEO Dario Amodei met with senior administration officials Friday amid a months-long public feud that saw the administration label Anthropic a "RADICAL LEFT, WOKE COMPANY" and national security threat. The discussions centered on potential collaboration around cybersecurity, AI leadership, and safety standards. The diplomatic thaw appears linked to Anthropic's newly unveiled Claude Mythos Preview, a cybersecurity-focused AI model that can reportedly identify vulnerabilities across major web browsers and operating systems. The model, described as Anthropic's most powerful to date, has already attracted corporate clients including Apple, Nvidia, and JPMorgan Chase. Government relations between Anthropic and the Pentagon deteriorated in February when the company refused to permit its technology for domestic surveillance or fully autonomous lethal weapons. The dispute led to Anthropic being designated a supply chain risk, a lawsuit challenging that designation, and public social media exchanges. Anthropic confirmed the White House meeting, stating it reflected the company's "ongoing commitment to engaging with the US government on the development of responsible AI." The company has reportedly briefed multiple government agencies on Mythos Preview's capabilities and recently retained Ballard Partners, a lob According to Axios, some intelligence community components and the Cybersecurity and Infrastructure Security Agency are already testing the cybersecurity model. A source familiar with negotiations suggested withholding the technology would be "grossly irresponsible" for U.S. national security and "a gift to China."