Security experts urge U.S. to lift Fable export ban

More than 100 cybersecurity leaders and researchers have signed an open letter calling on the U.S. to reverse an export restriction on Anthropic’s Fable 5 model. They argue the ban limits defenders more than attackers, since similar capabilities are available in other AI systems. The letter says the blocked jailbreak example should be treated as evidence of a vulnerability that can help improve defenses, rather than as proof that the model is uniquely dangerous. It also calls for AI regulation based on scientific evaluation, transparent enforcement, and a fair public process. The debate follows reporting that the restriction may have stemmed partly from a communication breakdown between Anthropic and U.S. officials, adding uncertainty around the policy’s original purpose. Supporters of the letter include security figures affiliated with companies and institutions such as Adobe, Zoom, Nvidia, Stanford HAI, and others.