Day 14 of the Fable 5 ban arrives with a deadline attached: by close of business today, the Commerce Department owes Congress an explanation for why Howard Lutnick’s office ordered Anthropic on Friday, June 13 to cut off access to Fable 5 and Mythos 5 “by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.” The letter, signed personally by the Commerce Secretary and addressed to CEO Dario Amodei, landed at 5:21pm ET, four days after Fable 5 shipped, and on the same afternoon SpaceX closed its public market debut at a $2.1 trillion valuation.
The official rationale was an export-control concern. The actual sequence, as reconstructed by the Wall Street Journal and amplified by CNBC and TechCrunch, was a phone call. Amazon CEO Andy Jassy contacted senior administration officials, including Treasury Secretary Scott Bessent, to flag a paper by Amazon security researchers describing a series of prompts that coaxed Fable 5 into surfacing information potentially useful for cyberattacks. Anthropic had received the paper privately and forwarded it to outside experts. Within days, the federal government had reclassified its roughly trillion-dollar competitor’s flagship models as a national-security matter.
Anthropic’s public statement is unusually pointed for a company in active litigation with the Pentagon. It says the only evidence presented was verbal, describing “a potential narrow, non-universal jailbreak,” and that the underlying capability is “widely available from other models, including OpenAI’s GPT-5.5.” Fable 5 was red-teamed for thousands of hours before launch by the U.S. government itself, the UK AI Safety Institute, and third parties. A person familiar with those reviews told CNBC that agencies had signed off on deployment without communicating any threat.
Katie Moussouris, the Luta Security founder who reviewed the Amazon paper, was blunter. The behavior “should never have triggered an export control.” By her account, the bypass amounted to asking the model to fix code rather than review it for security issues, and the capability couldn’t be patched without crippling the model’s defensive utility.
The political backdrop sharpens the picture. In February, President Trump ordered federal agencies to stop using Anthropic’s models after the company refused Pentagon contract terms covering autonomous weapons and domestic surveillance. In early March, the Pentagon designated Anthropic a “supply chain risk,” a label Anthropic is challenging in federal court. Former AI czar David Sacks has spent months accusing the company of “regulatory capture” and “fear-mongering.” OpenAI is reportedly preparing its own IPO filing.
What’s left is an export-control regime functioning as competitive policy. Daniel Remler of the Center for a New American Security put the structural point cleanly: “This sure looks mandatory if there are going to be consequences for not doing what the government says.” The 2008 TARP precedent established that “voluntary” federal requests carry the weight of orders when refusal invites retaliation. Two weeks in, the Fable 5 ban looks like that doctrine reaching the model layer, with a rival CEO holding the receiver.
Sources
- https://www.anthropic.com/news/fable-mythos-access
- https://techcrunch.com/2026/06/15/the-us-governments-anthropic-models-ban-was-never-about-an-ai-jailbreak/
- https://www.cnbc.com/2026/06/17/anthropic-ai-regulation-trump.html
- https://www.bloomberg.com/news/articles/2026-06-13/anthropic-says-us-limits-foreign-access-to-fable-5-mythos-5
- https://time.com/article/2026/06/13/anthropic-fable-mythos-ban-US-security/