Security

With AI, there’s never a dull moment. Anthropic released Claude Fable 5 and Mythos 5 on June 9, and four days later was forced to withdraw the models. US Commerce Secretary Howard Lutnick said in a letter to Anthropic CEO Dario Amodei that export of the models must be suspended, and it was feared that they could be used by military intelligence users in Russia or China, among other countries.

The letter arrived after 5pm, and the company was given 90 minutes to comply, with the government saying it had discovered a jailbreak in the model. Anthropic said it disagreed that the finding of a “narrow potential” jailbreak was cause for the recall, and if other frontier model providers were held to the same standard, no new models would ever see the light of day.