NewsBite analysis · AI, Security
The Decoder
Anthropic's Fable 5 is back worldwide after a two-week government ban over a jailbreak
Anthropic has announced the global reinstatement of its Fable 5 AI model after a two-week ban imposed by the US government due to identified security vulnerabilities related to jailbreaks. The model will be accessible through various enterprise platforms, with access restoration on major cloud services underway. As part of its response, Anthropic has enhanced safety measures to mitigate the risk of exploitation and is working with the government on security standards. The company emphasizes the need for industry-wide standards for managing jailbreak vulnerabilities and is collaborating with major tech firms to develop a robust security framework.
MonitorConfidence: HighImmediate horizon
Published 1 Jul 2026, 08:10 UTC3 min read
Visit original article ↗Lead image — The Decoder
Summary
TL;DR
- Fable 5 returns globally after a two-week US government ban.
- The ban was due to security findings related to jailbreak vulnerabilities.
- Anthropic is implementing an improved safety classification system.
- Collaborations with government partners will shape future AI model regulations.
- A new HackerOne program allows researchers to report AI-related security issues.
Interesting facts
- The jailbreak issue was related to Fable 5 identifying software vulnerabilities.
- The enhanced classifier will flag harmless requests more frequently.
- No uniform jailbreak solution exists for AI models, according to Anthropic.