Anthropic Says It’s Taking Claude Fable 5 Offline to Comply With US Government Order
Anthropic is disabling its Claude Fable 5 and Mythos 5 AI models due to a US government export control directive citing national security concerns, an unprecedented move amidst ongoing tensions with the Trump administration.
Intelligence analysis by Gemini 2.5 Flash
Anthropic has halted access to its new AI models, Claude Fable 5 and Mythos 5, after receiving a US government order citing national security concerns related to potential 'jailbreaking' methods. This incident escalates existing friction between Anthropic and the Trump administration, which previously designated Anthropic as a 'supply chain risk'.
Imagine a super-smart robot brain that helps with tricky computer problems. A company called Anthropic made a new version, but the grown-ups in charge of the country told them to switch it off because they worried someone might trick it into doing bad things. The company says the trick isn't a big deal, but they still turned it off to follow the rules.
Analysis
Anthropic announced it is taking its newly launched AI models, Claude Fable 5 and Mythos 5, offline following a directive from the US government. The company stated that it received an export control order on Friday afternoon, citing national security concerns. This move is unprecedented and marks a significant escalation in the relationship between Anthropic and the Trump administration.
The government's order specifically requested the suspension of access for "any foreign national, whether inside or outside the United States, including foreign national Anthropic employees." However, Anthropic chose to remove access for all customers to ensure full compliance.
Earlier this year, the Trump administration's Department of Defense labeled Anthropic a "supply chain risk" after the company attempted to establish guidelines on the military's use of its technology. This designation prevented government agencies and contractors from utilizing Anthropic's technology, prompting lawsuits from the AI company against the administration.
Claude Fable 5, a version of Anthropic's Mythos AI model, was publicly released earlier this week with built-in safeguards to prevent it from addressing topics like cybersecurity, biology, and chemistry. Anthropic had previously stated that this release was a collaboration with the US government. A preview version, Mythos Preview AI, had a limited rollout in April, intended to provide cybersecurity capabilities for organizations and mitigate potential misuse by malicious actors.
In a blog post, Anthropic indicated the government's letter arrived at 5:21 pm ET, without specific details regarding the national security concern. Anthropic believes the concern stems from a method of bypassing, or "jailbreaking," Fable 5. The company reviewed a demonstration of this technique, which reportedly identified a small number of minor, previously known vulnerabilities that other publicly available models could also discover without requiring a bypass. Anthropic maintains that its models have strong safeguards and that the identified jailbreak was narrow and would not significantly enhance an attacker's capabilities compared to other AI models. White House and Commerce Department spokespeople did not comment immediately.
Key points
- Anthropic is disabling Claude Fable 5 and Mythos 5 models under a US government export control order.
- The government order cites national security concerns, specifically a potential 'jailbreak' method for Fable 5.
- Anthropic claims the identified jailbreak is narrow and poses minor risks, comparable to other publicly available models.
- This incident adds to existing tensions between Anthropic and the Trump administration, which previously labeled Anthropic a 'supply chain risk'.
- Anthropic's CEO supports transparent government processes for blocking unsafe AI models, but argues this action did not adhere to those principles.
The incident could lead to clearer, more transparent government processes for regulating AI models, as advocated by Anthropic's CEO, fostering a safer development environment. It might also encourage AI companies to enhance their internal safeguards and collaborate more effectively with governments on national security concerns.
The immediate consequence is a setback for Anthropic, potentially stifling innovation and delaying the release of advanced AI models due to government intervention. This event could also create uncertainty for other AI developers regarding regulatory compliance and the future deployment of their technologies, potentially leading to a more restrictive AI landscape.
