President Trump said that a week earlier he might have considered Anthropic a national security threat, after his administration restricted access to the company's newest models.
June 19, 2026 · Trump Administration · Anthropic
"A Week Ago I Might Have Called Them a National Security Threat"
After ordering Anthropic to cut off foreign access to its newest models, Trump said he might have viewed the company as a national security threat — while adding he no longer sees it that way and does not want to shut it down.
2
models restricted —Fable 5 & Mythos 5
~150
signatories on an open letter urging reversal
100%
of users blocked — models taken fully offline to comply
An Escalating Dispute
Each step raised the stakes between the administration and Anthropic
Feb 2026
Pentagon labels firm a "supply chain risk"; halts Claude use
→
Mar 2026
Anthropic sues the Pentagon over the label
→
Jun 2026
Export controls force foreign access cut-off → models offline
The Directive (around June 13)
Block access to Fable 5 & Mythos 5 for all foreign users — citing jailbreak vulnerabilities flagged in Mythos-class models.
Scope: foreign nationals, foreigners residing in the U.S., and foreign-national employees
Response: Anthropic temporarily suspended the models for everyone
Critics warn
Could set a precedent that hampers defensive vulnerability research
Harder for AI firms to build tools that find and fix flaws
Signals U.S.-made AI may be unreliable — eroding foreign trust
Anthropic's stance
The government's stated concerns lack a basis
Has refused use of its models for domestic surveillance and autonomous weapons
Lifting conditions and the lawsuit are the next focus
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…