In a surprising diplomatic turn, the White House has held a high-level meeting with artificial intelligence company Anthropic, temporarily setting aside an ongoing legal battle to discuss the company's powerful new Claude Mythos model. The meeting signals a potential shift in the U.S. government's stance toward a firm it had previously criticized.
Anthropic CEO Dario Amodei met on Friday with Treasury Secretary Scott Bessent and White House Chief of Staff Susie Wiles, according to reports. The White House described the discussions as "productive and constructive," focusing on collaboration opportunities and safety protocols for advanced AI technology.
This meeting comes just one week after Anthropic released a preview of Claude Mythos, an AI tool the company claims can outperform humans in certain hacking and cybersecurity tasks. Researchers have described the model as "strikingly capable at computer security tasks," with the ability to find and exploit vulnerabilities in decades-old code.
"We discussed opportunities for collaboration, as well as shared approaches and protocols to address the challenges associated with scaling this technology," the White House stated, adding that the meeting had "explored the balance between advancing innovation and ensuring safety."
The diplomatic engagement marks a notable departure from recent tensions. Just two months ago, the White House had derided Anthropic as a "radical left, woke company," and former President Donald Trump had directed government agencies to stop using the firm's technology, calling its leadership "left wing nut jobs" on social media.
The backdrop to this meeting includes an ongoing legal battle. In March, Anthropic sued the Department of Defense and other federal agencies after receiving a "supply chain risk" designation—the first time a U.S. company has been publicly labeled as such. This designation indicates a technology is not secure enough for government use.
Anthropic argued in court that the designation was retaliation by Defense Secretary Pete Hegseth, stemming from the company's refusal to grant the Pentagon unrestricted access to its AI tools over concerns about potential misuse for mass domestic surveillance and fully autonomous weapons.
While a federal court in California has largely sided with Anthropic, an appeals court has denied the company's request to temporarily block the supply chain risk designation. Despite this, Anthropic's tools reportedly remain in use at all government agencies that had been using them before the designation.
Only a few dozen companies have been granted access to the Mythos preview so far. Last week, Amodei stated that Anthropic had "spoken to officials across the U.S. government" and offered to work with them on the technology.
When questioned by reporters about the White House meeting on Friday, Trump said he had "no idea" about the discussions, highlighting the unexpected nature of this diplomatic engagement between the administration and a company it had previously targeted.
The meeting suggests that despite political tensions, Anthropic's technology may be too strategically important for the U.S. government to ignore, particularly as concerns grow about cybersecurity capabilities in an increasingly digital world.