White House chief of staff Susie Wiles recently met with Anthropic CEO Dario Amodei to discuss the artificial intelligence company’s new Mythos model. This advanced AI model has garnered significant attention from the federal government due to its potential to transform national security and the economy. The White House, engaging with leading AI labs, underscored the need for a technical evaluation period for any new technology considered for federal use, emphasizing a productive and constructive discussion focused on collaboration and balancing innovation with safety. Anthropic stated that the meeting, which included senior administration officials, explored how the company and the U.S. government could partner on critical shared priorities such as cybersecurity, maintaining America's leadership in the AI race, and ensuring AI safety, expressing eagerness for ongoing dialogue.

This recent engagement follows a period of heightened tensions between the Trump administration and Anthropic, a company known for its safety-conscious approach and efforts to implement guardrails on AI development. Anthropic aims to minimize potential risks while maximizing AI's economic and national security benefits for the U.S. Previously, President Donald Trump attempted to prohibit all federal agencies from utilizing Anthropic’s chatbot Claude, citing a contract dispute with the Pentagon and asserting the administration would cease doing business with the company. When questioned about the Friday meeting, Trump stated he had “no idea.”

Further escalating the dispute, Defense Secretary Pete Hegseth sought to label Anthropic as a supply chain risk, an unusual move against a U.S. company that Anthropic subsequently challenged in two federal courts. Anthropic’s contention was rooted in its desire for assurance that the Pentagon would not deploy its technology in fully autonomous weapons or for the surveillance of American citizens. Hegseth, however, maintained that the company must permit any uses deemed lawful by the Pentagon. The judicial branch intervened when U.S. District Judge Rita Lin issued a ruling in March, effectively blocking the enforcement of Trump’s social media directive that ordered federal agencies to cease using Anthropic products.

Anthropic, upon announcing its Mythos model on April 7, described it as “strikingly capable,” leading the company to restrict its use to a select group of customers. This limitation stems from Mythos’s unparalleled ability to surpass human cybersecurity experts in identifying and exploiting computer vulnerabilities. While some industry observers initially questioned whether Anthropic’s claims of such powerful AI were merely a marketing tactic, even some of its most vocal critics have conceded that Mythos might indeed represent a significant leap forward in AI capabilities. David Sacks, an influential Anthropic critic and former White House AI and crypto czar, urged people to “take this seriously,” noting that “as the coding models become more and more capable, they are more capable at finding bugs… vulnerabilities… [and] stringing together multiple vulnerabilities and creating an exploit.”

The dual potential for significant benefits and considerable risks associated with the Mythos model has also captured international attention. The United Kingdom’s AI Security Institute evaluated the model, concluding it was a “step up” from previous iterations, which themselves were rapidly advancing. Their report highlighted that “Mythos Preview can exploit systems with weak security posture, and it is likely that more models with these capabilities will be developed.” Anthropic has also engaged in discussions with the European Union regarding its AI models, including advanced versions not yet released in Europe. In response to the model's profound implications for public safety, national security, and the economy, Anthropic has launched Project Glasswing, an initiative that unites major technology companies such as Amazon, Apple, Google, and Microsoft, alongside other firms like JPMorgan Chase, to collaboratively secure the world’s critical software infrastructure.

Jack Clark, Anthropic co-founder and policy chief, clarified at the Semafor World Economy conference that while Mythos is currently ahead of the curve, it is not a