Anthropic is turning away distillers' paying traffic

What do Anthropic's distillation defences cost legitimate users?

Anthropic says that in February it wrote of working to make distillation harder without affecting the experience for legitimate users, and that “We haven't met this goal completely.” Preserved thinking has required some developers to update their integrations, its classifiers can misidentify or block benign requests in fewer than 0.1% of cases, and it now charges users for those blocked requests. Since September 24 it has resumed charging for requests its safeguards block before Claude responds, in the specific areas, including distillation, where they have a false-positive rate under 0.1%.

Why does Anthropic treat distillation as more than a commercial problem?

It gives two reasons. It says distillation lets companies build models that approach the capability frontier but have, in practice, been released “without proportionally robust safeguards.” And it says illicitly distilled models “present a national security risk,” adding that the companies that distill US models are integrating the capabilities they gain “into foreign military, intelligence, and surveillance systems.”

What does Anthropic ask of other AI companies?

It says preventing illicit distillation “requires work and buy-in from all frontier model developers,” and that its own efforts would be undermined if attackers could easily access similarly capable models that lack similarly robust anti-distillation safeguards. Platforms, it says, can “materially slow down distillation” by quickly detecting and disrupting abuse, and by offboarding fraudulent accounts.