# Anthropic holds daily talks with Trump administration on AI security concerns > Source: > Published: 2026-06-18 14:06:39+00:00 # Anthropic holds daily talks with Trump administration on AI security concerns The Claude maker has been forced to disable its most advanced models for foreign users after a Commerce Department directive, and now its engineers are meeting daily with US officials to find a path forward. Anthropic’s technical staff have been sitting down with US government officials every single day since June 16 to resolve a standoff that saw the company’s most powerful AI models yanked from global access almost overnight. The trigger: a Commerce Department directive issued on June 12 ordering Anthropic to disable access to its Fable 5 and Mythos 5 models for foreign nationals, citing national security risks tied to potential jailbreak vulnerabilities that could enable cyberattacks. By June 13, the models were dark. ## How we got here The roots of this conflict stretch back months. In February 2026, the Trump administration instructed federal agencies to stop using Anthropic’s products entirely, a move tied to disputes over military applications of the company’s technology. Then came the executive order on June 2. President Trump signed a framework establishing a voluntary review process for national security evaluation of frontier AI models. Ten days later, the Commerce Department’s directive to Anthropic felt decidedly less optional. ## Banking-style regulation for AI Anthropic CEO Dario Amodei has called for banking-style checks on AI development: the same kind of stress tests, compliance requirements, and regulatory oversight that banks face after the 2008 financial crisis, applied to companies building frontier AI models. He’s also suggested that regulators should have the ability to pause development of potentially dangerous models entirely. ## The diplomatic dance President Trump weighed in on the negotiations while attending the G7 summit on June 17, saying that talks with Anthropic were progressing positively. No resolution has been reached as of mid-June. The daily meetings between Anthropic’s technical staff and US officials represent an unusual level of engagement between a private AI company and the federal government. The core issue revolves around jailbreak risks in Fable 5 and Mythos 5, with the government’s concern being that foreign nationals could exploit these vulnerabilities for offensive cyber operations. **Disclosure:** This article was edited by Editorial Team. For more information on how we create and review content, see our [Editorial Policy](https://cryptobriefing.com/editorial-policy/).