White House Chief of Staff Susie Wiles Holds Talks with Anthropic CEO on AI Model Mythos
White House Chief of Staff Susie Wiles engaged in discussions with Anthropic CEO Dario Amodei on Friday, focusing on the company's newly unveiled Mythos artificial intelligence model. This high-level meeting underscores the federal government's keen interest in how advanced AI technologies could reshape national security frameworks and drive economic innovation. A White House official, speaking anonymously to provide insights ahead of the formal announcement, confirmed that the administration is actively consulting with leading AI laboratories to assess their models and ensure software security protocols are robust.
The official emphasized that any novel technology considered for federal deployment would undergo a comprehensive technical evaluation period to mitigate potential risks. In a subsequent statement, the White House characterized the dialogue as productive and constructive, highlighting opportunities for collaborative efforts aimed at striking a balance between fostering innovation and upholding safety standards.
Anthropic's Perspective and Past Tensions with the Trump Administration
Anthropic released a statement detailing that Amodei's meeting included senior administration officials and explored avenues for cooperation between the San Francisco-based firm and the U.S. government. Key priorities discussed encompassed cybersecurity, maintaining America's competitive edge in the global AI race, and advancing AI safety measures. The company expressed optimism about continuing these discussions in the future.
This engagement occurs against a backdrop of heightened tensions between the safety-conscious Anthropic and the previous Trump administration. Anthropic has consistently advocated for implementing guardrails in AI development to minimize risks while maximizing economic and national security benefits for the United States. Former President Donald Trump previously attempted to halt all federal agencies from using Anthropic's chatbot Claude due to a contract dispute with the Pentagon, declaring in a February social media post that the administration would cease business with the company.
When questioned about the White House meeting while in Arizona, Trump stated he had no knowledge of the event. Additionally, Defense Secretary Pete Hegseth sought to label Anthropic as a supply chain risk, an unprecedented action against a domestic company that Anthropic has contested in two federal courts. The firm has demanded assurances that the Pentagon would not employ its technology in fully autonomous weapons or for surveilling American citizens, while Hegseth insisted on allowing any uses deemed lawful by the Pentagon.
Legal Developments and the Capabilities of the Mythos Model
In a significant legal ruling, U.S. District Judge Rita Lin blocked the enforcement of Trump's directive ordering federal agencies to discontinue using Anthropic products. Anthropic announced its Mythos model on April 7, describing it as strikingly capable to the extent that it is restricting access to select customers. The model reportedly surpasses human cybersecurity experts in identifying and exploiting computer vulnerabilities, raising both excitement and concerns within the industry.
While some industry experts have speculated whether Anthropic's warnings about overly powerful AI are merely marketing tactics, even critics acknowledge that Mythos may represent a substantial advancement in AI technology. David Sacks, a former White House AI and crypto czar and an influential critic, urged the public to take the developments seriously. On the All-In podcast, Sacks noted that as coding models become more sophisticated, their ability to detect bugs and vulnerabilities increases, potentially leading to more effective exploits.
International Attention and Collaborative Initiatives
The potential benefits and risks of the Mythos model have garnered international scrutiny. The United Kingdom's AI Security Institute evaluated the new model and reported it as a step up from previous iterations, which were already advancing rapidly. The institute warned that Mythos Preview could exploit systems with weak security postures, and similar capabilities are likely to emerge in future models.
Anthropic has also been in discussions with the European Union regarding its AI models, including advanced versions not yet released in Europe, as confirmed by European Commission spokesman Thomas Regnier. In conjunction with the Mythos announcement, Anthropic launched Project Glasswing, an initiative uniting tech giants like Amazon, Apple, Google, and Microsoft, along with firms such as JPMorgan Chase. This project aims to secure critical global software from severe fallout that the new model might pose to public safety, national security, and the economy.
Anthropic co-founder and policy chief Jack Clark, speaking at the Semafor World Economy conference, explained that Mythos is being released to a select group of major companies and organizations to help them identify vulnerabilities. Clark cautioned that while Mythos is currently ahead of the curve, it is not a unique model; he predicted that similar systems from other companies will emerge within months, and open-weight models from China with comparable capabilities could appear within a year to a year and a half. This progression underscores the urgent need for global preparedness in the face of increasingly powerful AI systems.
