Anthropic is closing in on a deal with the White House that would lift US restrictions on its two leading artificial intelligence models, according to sources familiar with the talks. The agreement, still taking shape, signals a potential thaw in government oversight of advanced AI systems. Anthropic PBC has been in active negotiations with the Trump administration to ease curbs on its flagship releases, Fable 5 and Mythos 5.
The move comes amid a broader debate over how tightly the US should regulate frontier AI technologies. Proponents argue that restrictions are needed to prevent misuse, while critics contend they stifle innovation and cede ground to foreign competitors. A deal could reshape the regulatory landscape for American AI firms.
Specific details of the proposed agreement remain sparse, as the talks are ongoing and confidential. Neither Anthropic nor the White House has publicly confirmed the negotiations. The Bloomberg report, citing unnamed sources, is the primary account of the progress.
If finalized, the arrangement would allow Anthropic to deploy its most advanced models without current limitations, potentially accelerating commercial applications in sectors from defense to healthcare. Other AI developers may seek similar accommodations, testing the administration's willingness to tailor rules.
Critics warn that lifting restrictions prematurely could undermine safety standards, especially as AI capabilities grow rapidly. The counter argument emphasizes that without robust oversight, models like Fable 5 and Mythos 5 might be used for disinformation or automated cyberattacks.