The Anthropic ‘Fable’ saga proves: we have opened the AI Pandora’s box. What now? | Nathan E Sanders and Bruce Schneier
TL;DR
Anthropic released Fable on 9 June as a constrained version of Mythos. Three days later, the US government treated it as a dangerous munition and used export controls to block foreign access. Anthropic could not reliably separate US users from non-US users, so it shut access down for everyone. The move exposes how blunt national controls are when AI services are globally networked.
Nauti's Take
Locking Fable away as if one model were the weapon sounds decisive, but it targets the wrong layer. Once agents can turn vague goals into actions across real systems, blacklists and export controls mostly buy time.
That time should go into hard evaluation duties, traceable tool permissions, operator liability and public safety research. Anthropic should not hide behind PR fog, but blaming one company is too small a diagnosis.
Briefingshow
The Fable episode shows that AI risk no longer sits inside one product release. Even if a government slows one model, capability can move through smaller models, toolchains and open-source harnesses. Serious governance has to cover access, deployment context, liability, audits and infrastructure, not just model labels.
Sources