Anthropic Shuts Down Models Amidst Security Tensions with Trump Administration

Following a directive from the US Commerce Department, Anthropic has suspended its newly released Fable 5 and Mythos 5 models due to national security concerns. The administration is reportedly addressing a jailbreak vulnerability that bypasses safeguards on cybersecurity and biological data. While Anthropic has complied with the mandate, it maintains that the vulnerabilities are minor and not unique to its systems. A more comprehensive explanation from the company is expected within 24 hours.
In a significant escalation of government oversight, Anthropic has abruptly terminated access to its recently launched Fable 5 and Mythos 5 models. This decision follows a formal directive from the US Commerce Department, which has subjected the new models to rigorous export controls. Anthropic indicated that disabling these models for all customers was the only immediate means of ensuring compliance with the administrative mandate. Central to the government's intervention is a purported jailbreak capable of circumventing classifier-based safeguards. These protections were specifically engineered to prevent users from generating content related to cybersecurity, chemistry, and biology. An administration official indicated that the suspension provides essential time for the national security apparatus to be hardened against such vulnerabilities—a process anticipated to conclude within several weeks. Despite its compliance, Anthropic has expressed profound disagreement with the government’s stance. The company revealed that the administration had only provided verbal evidence of a narrow, non-universal jailbreak involving the review of codebase for minor software vulnerabilities. Anthropic maintains that these capabilities are not unique to its models, citing GPT-5.5 as having comparable functionality. They further cautioned that applying such a stringent standard could potentially paralyze the deployment of frontier models across the entire AI industry. This directive arrives on the heels of an executive order signed by President Trump, which encourages AI manufacturers to undergo voluntary security assessments. Anthropic has categorized the resulting disruption as a misunderstanding and has offered an apology to its global user base. While the company intends to release further details within 24 hours, the immediate removal of Fable 5 and Mythos 5 remains in effect. Notably, access to Anthropic’s other existing models remains unaffected, though the incident has ignited a broader debate regarding the intersection of commercial innovation and national security requirements.
Take a position. Out loud, if you can.
Four ways to start. Pick one and try saying it before you scroll on.
Tip · Record yourself, use in a notebook, or practice with a language partner.
What triggered the shutdown of the models?
Complex Multi-Clause Subordination
Advanced sentence structures use multiple subordinate clauses to provide context, contrast, and precision.
“Anthropic indicated that disabling these models for all customers was the only immediate means of ensuring compliance with the administrative mandate.”
What to know · C1
Try saying this aloud
Scenario: Analyzing geopolitical impact on corporate strategy.
- 01“This marks a significant escalation in oversight.”
- 02“We are operating at the intersection of innovation and security.”
- 03“Such a stringent standard could paralyze the industry.”
Register tip · formal
🔑Key Phrases
Describes an increase in the state's control over private business.
The directive represents an escalation of government oversight in tech.
Refers to finding ways around specific technical safety mechanisms.
Hackers are constantly circumventing classifier-based safeguards.
Implementing very strict requirements or rules.
Applying such a stringent standard would change the market.
The point where business growth and new technology meet.
The policy sits at the intersection of commercial innovation and public safety.
🎙️ Article Audio — Kokoro TTS
Anthropic Shuts Down Models Amidst Security Tensions with Trump Administration
Adapted from Ars Technica · Read the original. LectoPress rewrites the facts as original graded-reader text for language learners.
Get stories at your level, every day
C1 · EN · delivered to your inbox · unsubscribe any time
Customize language, level & topics → full preferences


