Artificial intelligence company Anthropic has announced that it will restore public access to its advanced AI models, Claude Fable 5 and Mythos 5, after U.S. authorities lifted export restrictions that had forced the company to suspend the models earlier this month.
The decision follows discussions between Anthropic and U.S. government officials, with the company introducing additional security measures designed to address concerns about potential misuse of its AI technology.
Advanced AI Models Return With Stronger Safeguards
According to Anthropic, Claude Fable 5 and Mythos 5 are being brought back online with upgraded safety systems.
The company said the latest version includes new AI classifiers capable of detecting and blocking a broader range of cybersecurity-related requests. These improvements are intended to reduce the risk of the models being exploited through jailbreak techniques or other attempts to bypass built-in safety protections.
Anthropic explained that the updated safeguards were developed following productive discussions with U.S. authorities.
Why Were the Models Taken Offline?
Anthropic suspended both models on June 12 after receiving a U.S. government export control directive.
The order required the company to restrict access to the AI models for all foreign nationals, including foreign employees working within Anthropic. Rather than creating separate access rules, the company temporarily disabled both models for all users to ensure full compliance.
The restrictions came after officials raised concerns over reports that researchers had discovered a jailbreak capable of making Claude Fable 5 identify software vulnerabilities.
Government Clears Public Redeployment
Following a review process, U.S. officials approved the return of the AI models.
U.S. Secretary of Commerce Howard Lutnick said government agencies had worked closely with Anthropic over the past two weeks to evaluate the models while ensuring they met national security requirements.
White House Chief of Staff Susie Wiles also emphasized that the administration’s goal is to support the deployment of advanced AI technology while maintaining appropriate safety standards.
Anthropic Disagreed With the Original Restrictions
Anthropic has consistently argued that the reported jailbreak did not justify removing the models from public use.
The company stated that the vulnerability represented only a limited method of bypassing safeguards rather than a universal exploit capable of disabling all safety systems.
According to Anthropic, less advanced AI models could produce similar cybersecurity-related responses, meaning the issue was not unique to Claude Fable 5.
The company also warned that recalling frontier AI models over isolated jailbreaks could slow innovation across the entire artificial intelligence industry if applied as a general regulatory standard.
Expanded Cooperation With the U.S. Government
Alongside the relaunch, Anthropic announced that it will deepen its collaboration with U.S. government agencies on AI safety.
The partnership will include:
- Early government access to advanced AI models before public release
- Joint testing of model safeguards
- Information sharing on jailbreak techniques and misuse
- Collaborative cybersecurity research
- Ongoing evaluation of frontier AI systems
Anthropic said these efforts are designed to improve the safety and reliability of increasingly powerful AI technologies.
Industry Collaboration on AI Security
The company is also working with major technology firms, including Amazon, Microsoft, and Google, through Project Glasswing, a cybersecurity initiative launched earlier this year.
The project aims to develop a common framework for evaluating AI jailbreaks and classifying the severity of security risks across advanced AI models.
By creating shared standards, participating companies hope to improve how the industry responds to emerging AI security challenges.
Calls for Stronger AI Regulation Continue
The restoration of Claude Fable 5 comes shortly after Anthropic released its policy proposal calling for stricter oversight of frontier AI systems.
The company has advocated for:
- Independent model evaluations
- Mandatory safety testing
- Stronger cybersecurity requirements
- Clear enforcement mechanisms
- Greater transparency around advanced AI capabilities
Anthropic believes these measures will help reduce risks associated with powerful AI systems while supporting responsible innovation.
The Bottom Line
With U.S. authorities lifting export restrictions, Anthropic is moving forward with the public relaunch of Claude Fable 5 and Mythos 5, this time with enhanced security protections and closer cooperation with government agencies.
The episode highlights the growing challenge of balancing rapid AI innovation with national security concerns. As governments and AI companies continue developing new safety standards, the outcome could shape how advanced artificial intelligence is deployed and regulated worldwide.
Disclaimer: This article is for informational purposes only and should not be considered legal, financial, or investment advice.
