Trump says he didn't know Anthropic's CEO was meeting with White House officials

Anthropic CEO Dario Amodei met with senior White House officials on Friday, marking a surprising turn in the company's relationship with the Trump administration. The meeting, confirmed by multiple sources, centered on Mythos Preview - a new AI model designed for cybersecurity applications. It comes less than two months after the administration banned federal use of Anthropic's flagship AI, Claude, over concerns about data handling and foreign influence.

Despite the high-level discussion, President Donald Trump told reporters he had no idea the meeting was taking place. 'I didn't know about it, and I'm not happy about that,' Trump said during a brief exchange outside the Oval Office. The comment underscores potential rifts within the administration over how to handle fast-evolving AI technologies and which companies should have access to government partnerships.

White House officials, however, pushed back on any suggestion of disarray. A spokesperson called the meeting 'productive and constructive' and emphasized that exploring emerging technologies is part of the administration's broader strategy to maintain U.S. leadership in AI. They noted that discussions focused strictly on national security applications and did not constitute a formal policy shift.

Still, the timing raises eyebrows. The ban on Claude was framed as a national security move, particularly due to concerns about cloud infrastructure and data routing. Now, engaging with the same company on a different - but equally sensitive - AI product suggests the administration may be drawing finer distinctions between AI use cases, or that internal pressure is mounting to reassess hardline positions.

Resistance appears to be coming from within the Pentagon, where some officials remain wary of Anthropic's access to government systems. According to reports, defense leaders worry that even limited collaboration could create backdoor risks, especially if Mythos Preview integrates with military or intelligence networks. That tension highlights the challenge of balancing innovation with security in an era where AI capabilities are advancing faster than policy.

The episode also shines a light on the broader uncertainty around the administration's AI strategy. With major elections on the horizon and global competitors pushing forward, decisions about which companies to trust - and when to reverse course - carry weight beyond any single meeting. For Anthropic, the White House discussion may signal a path toward rehabilitation, but only if it can navigate the political and bureaucratic hurdles still in place.

For now, the mixed messages - a CEO welcomed at the White House while the president claims ignorance - leave the tech industry and policymakers watching closely. Whether this is the start of a recalibration or just a misstep in communication, it's clear that AI is becoming one of the most politically charged technologies of the moment.