Anthropic meets White House after pulling AI models over access dispute
Anthropic executives met senior US officials following the company’s decision to block public access to its newest AI systems rather than comply with a government order to bar foreign nationals from using them. Anthropic withdrew both its Fable 5 model, the publicly available version, and Mythos 5, a more capable variant offered to a limited set of organizations, after the White House demanded nationality-based restrictions on short notice.
The underlying concern is foreign access to frontier models, with China cited as the specific worry. Rather than build and enforce per-user nationality screening, Anthropic chose to pull the releases entirely — a move that forced the dispute into direct talks with the administration.
The episode is drawing criticism that the US is regulating advanced AI case by case rather than through clear rules. Without a published standard for what access controls a frontier developer must apply, companies face demands that can arrive abruptly and differ from one product to the next, making compliance unpredictable and pushing high-stakes decisions into closed-door negotiations.