White House and Anthropic hold ‘productive’ meeting amid fears over Mythos model

AI firm’s CEO meets with US officials following legal dispute

The White House described a meeting with Dario Amodei, the head of artificial intelligence company Anthropic, as “fruitful and cooperative.” This encounter occurred a week after Anthropic unveiled its Claude Mythos preview, an AI system claiming to surpass human capabilities in specific cybersecurity and hacking functions. Amodei engaged in discussions with Treasury Secretary Scott Bessent and White House Chief of Staff Susie Wiles, according to Axios. A spokesperson for Anthropic declined to comment on the session, which took place two months after the administration had labeled the firm a “radical left, woke company.”

Mythos model’s capabilities draw attention

Anthropic’s Mythos tool has attracted significant interest for its alleged proficiency in identifying hidden flaws in legacy code and autonomously exploiting them. Researchers have called its performance in cybersecurity tasks “remarkably advanced.” Despite the White House’s earlier criticism, the meeting suggests the government may now recognize the tool’s importance. The White House stated it explored ways to balance innovation with safety, highlighting collaboration opportunities with Anthropic.

Legal battle and supply chain risks

In March, Anthropic filed a lawsuit against the Department of Defense and other federal agencies after being tagged as a “supply chain risk.” This marked the first time a U.S. company received such a designation, implying its technology was deemed insecure for government applications. Despite the label, Anthropic’s tools continue to be utilized by agencies that previously relied on them. The firm argued in court that the designation was a retaliatory move by Defense Secretary Pete Hegseth, citing concerns over mass surveillance and autonomous weapons.

Trump’s stance and recent shift

Until Friday, the White House had expressed little praise for Anthropic, aligning with Trump’s directive to halt its use across government agencies. Trump had previously mocked the company on social media, calling its leadership “left wing nut jobs” and accusing them of “strong arming” the defense sector. When asked about the meeting during an Arizona event, Trump admitted he was unaware of Amodei’s visit, signaling a potential shift in his stance.

“We discussed opportunities for collaboration, as well as shared approaches and protocols to address the challenges associated with scaling this technology,” the White House said.

Anthropic’s ongoing involvement in high-level government and military projects since 2024 underscores its strategic value. While a California federal court largely supported the company’s claims, an appeals court rejected its request to temporarily suspend the supply chain risk designation. This development highlights the tension between innovation and security in the evolving AI landscape.