Export controls are an AI kill switch, the process needs reform

Export controls are an AI kill switch, the process needs reform, according to the article’s account of U.S. actions involving leading frontier models. It says a June 12 export control order tied to Anthropic forced the company to withdraw its most powerful models, Mythos 5 and Fable 5, from public availability because Anthropic could not reliably filter access by nationality or citizenship. The piece also notes OpenAI limited release of GPT-5.6 at the administration’s request, with later partial changes: on June 30, the administration told Anthropic that a license would no longer be required for Mythos or Fable, and on July 8 it approved a “wide release” of GPT-5.6. The article argues the approach functioned like a de facto licensing regime, raising concerns about due process and future uncertainty.