Anthropic seeks agreement with Trump officials to restore access to its powerful models

According to reports, negotiations between the company Anthropic and officials from the Trump administration were still ongoing on Monday, but they failed to reach an agreement on resolving security concerns. These concerns prompted the White House to restrict access to the latest models of the artificial intelligence company, and both sides are eager to find a solution. Last Friday, after the Trump administration announced that foreign governments, businesses, and many individuals were banned from accessing the Fable 5 model, Anthropic closed access for all users to comply with the new regulations. The company also took the same measures for the similarly restricted stronger model, Mythos 5. According to sources, Monday's latest discussions about Fable were led by the U.S. Department of Commerce and the U.S. National Cyber Director, Shawn Kane-Cros, with senior technical staff from Anthropic also participating in the discussions. "Since the government first reached out, we have held numerous online meetings. Both sides are moving quickly to resolve this matter," a spokesman for Anthropic said in the company's first public comments since last Friday, "This is part of our ongoing collaboration with the government to protect critical U.S. infrastructure and maintain America's leading position in the field of cyber defense." According to sources, some government officials have indicated that the resolution should include Anthropic acknowledging that they could have done better in launching Fable and communicating with the White House. Anthropic stated on Friday that the security vulnerabilities they were made aware of seemed relatively straightforward, and they had extensive cooperation with the government before releasing the model. Monday's discussions included Anthropic's Chief Security Researcher Nicholas Kalini, Logan Graham, who is in charge of the model risk assessment team, and the company's security chief, Dave Ohr.