A group of cybersecurity experts is pressing the White House to remove export-control restrictions on Anthropic’s most powerful models, Fable and Mythos. The group argues that limiting access could make it harder for defenders to secure software and products.
The dispute highlights a recurring tension in frontier AI policy: the same models that raise national-security concerns can also improve defensive security work. Blanket restrictions may reduce misuse risk, but they can also slow legitimate users.
The debate is likely to matter for other model providers as governments try to draw lines around advanced capabilities without cutting off trusted security teams.