U.S. Secretary of State Marco Rubio has circulated guidance telling American diplomats to counter international narratives that portray the United States as possessing a unilateral "kill switch" over its technology exports, according to a July 16 State Department cable. The memo, distributed globally, is intended to blunt criticism triggered by a recent U.S. government action that temporarily limited access to the nation’s most advanced artificial intelligence models.
The cable comes in the aftermath of a short-lived decision by the administration on June 12 to block foreign nationals from accessing two advanced models from Anthropic, named Mythos and Fable, on national security grounds. That restriction led the AI company to suspend access to those models worldwide. Although the ban was lifted later in June, the episode helped spark international debate over the reliability of U.S. technology availability for allies and partners.
In the memo, diplomats were instructed to push back on a narrative that has taken hold among some foreign lawmakers and industry figures - namely, that U.S. actions amount to an ability to flick a switch and cut off critical technologies. The cable emphasizes that temporary measures such as pausing narrow uses of technology or requesting a 30-day cybersecurity testing window prior to release do not constitute a "Kill Switch." It states that there is no single government "magic button" and characterizes the portrayal as exaggerated and lacking nuance.
The cable also urged U.S. diplomats to counter what it called "so-called 'digital sovereignty'" initiatives. It described these initiatives as attempts by foreign governments to limit access by American technology firms to their markets, impose rules requiring local data or infrastructure, levy "network usage fees," or mandate adherence to local content moderation and other regulations. The guidance instructed diplomats to argue that such measures could push national technology strategies toward inefficient, locally built solutions rather than leveraging established American products.
Part of the suggested messaging framed American AI offerings as the superior choice. The cable advised diplomats to highlight that American AI firms are capable of constructing large, independent AI infrastructure with secure, resilient supply chains that reduce the risk of hidden access points. "They build it. It’s yours," the document instructs representatives to say when describing the benefits of American AI systems.
The guidance was circulated amid broader U.S. steps to bring more oversight to AI releases. On June 2, the administration issued an executive order requesting that AI companies voluntarily submit their models for 30 days of cybersecurity testing prior to release. Taken together with the June 12 action on Anthropic, critics said the moves signaled a new willingness on the part of the U.S. government to regulate and, if necessary, restrict the distribution of AI technologies, potentially even to allied countries.
That perception has drawn criticism in Europe and praise for local alternatives in parts of Asia. Some European lawmakers have voiced concern that dependence on U.S. frontier AI models could leave their economies vulnerable to sudden restrictions. French deputy Christophe Grudler is quoted in criticism that the events demonstrated the United States holds a genuine "kill-switch" over crucial technologies and appears willing to use it. Aura Salla, a member of the European People’s Party, expressed worry that Europe could not continue to rely on a tech stack over which access might be switched off overnight by a foreign government.
Edward Fishman, director of the Maurice R. Greenberg Center for Geoeconomics at the Council on Foreign Relations, said he understood the safety rationale behind the initial restrictions on Anthropic’s models, which are capable of greatly amplifying complex cyber operations. At the same time, he observed that the U.S. administration’s prior use of coercive economic tools, including tariffs and sanctions, complicates attempts to reassure foreign partners. European concerns, he noted, center on whether heavy reliance on U.S. frontier models could, at some point, be turned into an economic lever against them.
Salla told diplomats and policymakers that recurrent threats and pressure from the administration have produced a climate of distrust. She said her government and others have not seen convincing evidence that the United States will be a consistently reliable partner on technology access, and that rhetoric alone will be insufficient to rebuild confidence.
Context and diplomatic aim
The cable represents a diplomatic effort to reframe recent U.S. actions in the AI space as precautionary safety steps rather than an instrument of geopolitical control. It instructs embassy and consulate staff to emphasize nuanced explanations for temporary restrictions and voluntary testing timelines, while arguing against policy shifts that would prioritize local alternatives over American products.
Communication strategy recommended
- Describe short pauses or voluntary testing as safety-oriented rather than coercive.
- Reject simplistic claims about a single government "kill switch."
- Promote American AI infrastructure as secure, independent, and preferable to newly built local substitutes.
The memo does not single out any company by name in its guidance, nor does it explicitly reference the June 12 action or the June 2 executive order in the talking points section. It nonetheless appears to be a direct response to the international reaction that followed those events.
The State Department framing attempts to strike a balance between acknowledging security-driven controls and defending U.S. commercial leadership in AI. Whether that framing will assuage concerns among European lawmakers and industry leaders who experienced the brief suspension of access remains unclear, as officials and observers point to lingering mistrust.