Anthropic is challenging a U.S. government order that bars foreign nationals from using CloudFable 5 with the support of several cybersecurity professionals and export control experts.
The company met with White House officials in Washington this week to try to reverse or limit the measures announced Friday.
Following the order, Anthropic shut down Fable 5 entirely, citing the difficulty of filtering out non-US citizens in real time. The company is now working with the government to find a solution.
Background and why cybersecurity experts are supporting Anthropic
Anthropic released Fable 5 as a consumer-facing version of its Mythos-class model architecture, with guardrails in place to prevent sensitive questions related to cybersecurity, biology, and chemistry.
This model was made available for free to Pro, Max and Enterprise users for a limited period ending on June 22.
Concerns about potential jailbreak vulnerabilities arose from Amazon’s cybersecurity research team.
The issue was escalated to the White House by Amazon’s CEO, leading the US government to ban foreign access to the model. In response, Anthropic took the Fable 5 offline for all users.
This is not the first time Anthropic has faced tensions with the US government. In March, the Pentagon labeled Anthropic as a supply chain risk and instructed federal agencies to stop using its models after the company refused to allow its products to be used for autonomous weapons or large-scale domestic surveillance.
A federal judge has ruled that the Pentagon’s directive cannot be enforced while Anthropic’s lawsuit against the designation continues.
Dozens of cybersecurity leaders have signed an open letter urging the US government to lift the ban on Fable 5. The letter said Anthropic “has built numerous protections into the Fable model to prevent its use for cyber offensive purposes.”
The signers argue that the jailbreak techniques cited by White House officials apply equally to other publicly available models, meaning that banning the Fable 5 does not specifically address the core issue.
The letter warns that removing a capable model like Fable from cybersecurity operations “while U.S. adversaries are rapidly advancing is dangerous.”
Export control experts have also questioned whether the US government has the legal authority to impose the ban in its current form. Their objections add a legal dimension to the dispute that now includes technical, security and regulatory questions.
The US government’s position on Fable 5 so far
The US government has not publicly stated the specific concerns behind the sanctions. Reuters reports that officials are seeking assurances from Anthropic that the model will not be used to harm US citizens.
Commerce Secretary Howard Lutnick participated in the discussions and is expected to continue talks with Anthropic CEO Dario Amodei during the G7 meetings in France. National cyber director Sean Cairncross also attended the meetings.
Anthropic’s earlier review of the cited jailbreak found that it exposed a small number of minor vulnerabilities that are already known and could be discovered by other publicly available models without bypass techniques.
What does this suspension mean for users, and why it matters
The Fable 5 and Mythos 5 are currently offline for all users as negotiations are ongoing. Anthropic has not provided any estimated timeline for his return.
Pro, Max, and Enterprise users who activated the Fable 5 during the free preview are unable to access the service at this time. Other cloud models, such as Opus 4.8, remain available and are not affected by these restrictions.
The controversy is becoming a prime example of how aggressively governments can regulate cutting-edge AI systems. The outcome could affect how AI companies handle export controls, security reviews and government access requests in the future.
It could also determine whether product-level restrictions or blanket regulations would apply to the entire industry due to national security concerns.
Anthropic hasn’t provided a timeline for when the Fable 5 might be available again. Upcoming G7 meetings in France may offer the next public update on the talks.
Thanks for being a Ghax reader. The post Anthropic Pushes Back Against US Order Banning Cloudflare 5 Backed by Cybersecurity Experts appeared first on gHacks.
