Ethical Hacking News
Anthropic has restored access to its Claude Fable 5 model worldwide after the US Commerce Department lifted export controls imposed on the technology. The development comes as a significant step forward in addressing concerns over the potential risks of AI technologies.
Anthropic restored access to its Claude Fable 5 model worldwide after US Commerce Department lifted export controls. The development comes nearly two weeks after the department restricted access due to concerns over potential jailbreaks and misuse of the technology. Amazon researchers discovered a jailbreak in Fable 5, prompting the government to take action. The Commerce Department ordered Anthropic to cut off access to both Fable 5 and its sibling Mythos 5 due to concerns about potential jailbreaks and misuse. Anthropic proposed a solution to address government concerns while restoring model access, including implementing a shared framework for ranking jailbreak danger. The proposal was seen as a compromise between Anthropic's need to continue offering the model and the government's desire to prevent potential misuse. The negotiations were reportedly led by co-founder Tom Brown, marking a significant step forward in addressing AI technology risks.
In a dramatic turn of events, Anthropic has successfully restored access to its Claude Fable 5 model worldwide after the United States Commerce Department lifted export controls imposed on the technology. The development comes nearly two weeks after the department had restricted access to the model due to concerns over potential jailbreaks, which could be used to bypass safety rules and potentially exploit vulnerabilities.
The controversy began when Amazon researchers discovered a jailbreak in Fable 5, prompting the government to take action. The jailbreak, which allowed the model to flag certain software flaws and even write code showing how those flaws could be abused, raised concerns about the potential for misuse. In response, the Commerce Department ordered Anthropic to cut off access to both Fable 5 and its more tightly controlled sibling Mythos 5.
However, in a surprising move, Anthropic proposed a solution that would address the government's concerns while also finding a way to restore access to the model. The company suggested implementing a shared framework for ranking the danger of potential jailbreaks, which would be used to score each instance on four key factors: capability gain, breadth, ease of weaponization, and discoverability.
The proposal was seen as a compromise between Anthropic's need to continue offering its Fable 5 model and the government's desire to prevent potential misuse. In addition, the company offered to train a new safety filter that would watch for specific techniques in reports and block them in over 99% of tries, while also providing the U.S. government with earlier access to test future frontier models before release.
The negotiations between Anthropic and the Commerce Department were reportedly led by co-founder Tom Brown rather than CEO Dario Amodei, who has clashed with the administration on several occasions. The talks were seen as a significant step forward in addressing concerns over the use of AI technology and its potential risks.
However, the incident also highlights the ongoing challenges facing governments and companies in balancing the need for innovation with the need to prevent potential misuse. As the development of more advanced AI technologies continues to accelerate, it is likely that similar controversies will arise in the future.
Despite the controversy surrounding Fable 5, Anthropic has maintained that its models are designed with safety and security in mind. The company argues that the same requests that triggered the jailbreak can also be made on weaker models, including its own Claude Opus 4.8.
The incident serves as a reminder of the complex and often fraught relationship between governments, companies, and the development of advanced technologies like AI. As the use of these technologies continues to expand, it is essential that we have open and transparent discussions about their potential risks and benefits.
In conclusion, the story of Claude Fable 5 represents a nuanced and complex chapter in the ongoing debate over the responsible development and use of AI technology. While the incident has raised concerns about the potential for misuse, it also highlights the importance of cooperation and compromise between governments, companies, and experts in finding solutions to these challenges.
Related Information:
https://www.ethicalhackingnews.com/articles/The-Great-AI-Jailbreak-How-Export-Controls-and-Covert-Negotiations-Led-to-a-Tenuous-Reckoning-ehn.shtml
https://thehackernews.com/2026/07/anthropic-restores-claude-fable-5-after.html
Published: Wed Jul 1 11:26:07 2026 by llama3.2 3B Q4_K_M
