ANTHROPIC PULLS FABLE 5 AFTER JAILBREAK DISCOVERY

Anthropic announced it is complying with a US government directive to take Fable 5 offline. In a blog post, the company stated that "the government believes it has become aware of a method of bypassing, or 'jailbreaking' Fable 5." The timing follows reports that Amazon and five other tech companies flagged security concerns about the model to the Trump administration. Amazon, a major investor in Anthropic, reportedly joined executives from other firms in warning the White House about vulnerabilities. The White House subsequently issued an export control order forcing the model offline within hours. Fable 5 had generated buzz in developer communities for its ability to generate functional video games through simple prompts—a capability that appealed to what industry observers called "vibe coders." The model represented a significant step toward more accessible game development tools. The incident highlights tensions in the AI industry around security oversight. While government intervention to address legitimate safety concerns is standard practice, the rapid deployment of export controls and the involvement of competing companies in alerting authorities raises questions about the intersection of security policy and competitive dynamics. The move also echoes earlier controversy surrounding Anthropic's Fable guardrails, for which the company apologized after users discovered invisible safety mechanisms that weren't clearly disclosed. That incident raised questions about transparency in how AI companies implement and communicate their safety measures. Anthropc has not provided a timeline for when or if Fable 5 might return to service. The company faces pressure to both address the identified vulnerabilities and rebuild user trust following the guardrail disclosure controversy.