Anthropic floats proposal to Lutnick to end US ban of powerful ‘Mythos,’ ‘Fable’ AI models: sources

1 1 minute read

Anthropic executives have made a commitment to collaborate more closely with the White House in a bid to address security concerns surrounding its powerful AI models, “Mythos” and “Fable,” as reported by The Post.

The ongoing discussions include a promise from Anthropic to enhance communication with the Trump administration and promptly resolve any security issues moving forward, according to a source familiar with the matter on Thursday.

Although talks between Anthropic and Trump officials are progressing positively, an exact timeline for a permanent solution remains uncertain, the source added.

Anthropic CEO Dario Amodei pictured at G7 Summit in France. Dominique Jacovides-Pool/SIPA/Shutterstock

As of Thursday, Anthropic cofounder Tom Brown and Sarah Heck are leading negotiations with Trump administration officials, as stated by a second source with insight into the situation.

Requests for comments from Anthropic and the Commerce Department were not immediately returned.

Last week, Anthropic sent senior officials to Washington DC after export controls were imposed on “Mythos” and “Fable” by the Trump administration.

The crackdown on Anthropic’s models followed a warning from Amazon CEO Andy Jassy about potential safety vulnerabilities that could be exploited. Anthropic decided to remove the models entirely to comply with the export controls.

The issue was discussed at the G7 Summit in France, with President Trump mentioning positive progress in talks with Anthropic, and Dario Amodei urging unity in AI regulation approaches among world leaders.

Commerce Sec. Howard Lutnick slapped export controls on Anthropic's AI models. — Commerce Sec. Howard Lutnick slapped export controls on Anthropic’s AI models. YOAN VALAT/EPA/Shutterstock

An Anthropic spokesperson stated earlier this week that both parties are working swiftly to resolve the situation, emphasizing the company’s commitment to safeguarding US critical infrastructure and cyber defense.

Coincidentally, the White House’s crackdown on Anthropic followed public remarks by Amodei advocating for government intervention in shutting down risky AI models, despite downplaying safety concerns as a minor issue.

According to a senior administration official, the White House had been alerted by multiple companies, including Amazon, about potential security risks before taking action.

“Had Anthropic taken the matter seriously and addressed it promptly instead of dismissing it, this situation could have been avoided,” the official stated.