Anthropic and Trump Administration Ease AI Rift as NSA Deploys Mythos in Cyber Operations
Updated
Updated · Gizmodo · Jun 5
Anthropic and Trump Administration Ease AI Rift as NSA Deploys Mythos in Cyber Operations
1 articles · Updated · Gizmodo · Jun 5
Summary
Reuters reported Friday that tensions between Anthropic and the Trump administration are easing, softening a clash that had threatened the company’s government business ahead of a planned IPO.
The dispute flared after Anthropic refused to loosen Claude’s guardrails for autonomous weapons and mass surveillance of Americans, prompting the Pentagon in March to label it a supply-chain risk and start purging its tools.
About 6 Anthropic engineers are now installed at the NSA, the Financial Times reported, customizing the company’s Mythos model as the agency uses it for offensive cyber operations likely aimed at China and Iran.
The thaw suggests national security agencies still want access to top-tier AI despite Defense Secretary Pete Hegseth’s earlier push to blacklist Anthropic after talks with CEO Dario Amodei broke down.
The shift also unfolds amid a broader White House scramble over AI influence, with rivals such as OpenAI and xAI courting Trump and OpenAI reportedly floating a possible U.S. government stake.