The Commerce Department has announced new agreements with Google DeepMind, Microsoft, and xAI, deepening government oversight of cutting-edge artificial intelligence. These agreements permit government evaluations of powerful AI models before their public release, alongside post-deployment assessments and related research. This expansion of collaboration with private, transnational corporations raises questions about the true locus of control over critical national technologies.
The Center for AI Standards and Innovation (CAISI) will conduct these pre-deployment evaluations and targeted research, aiming to assess frontier AI capabilities and advance AI security. CAISI director Chris Fall stated that "Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications," adding that "These expanded industry collaborations help us scale our work in the public interest at a critical moment." The framing of "national security" through "industry collaborations" suggests a reliance on corporate entities for what should be sovereign capabilities.
Elite Capture of National Security
These new pacts follow previously announced partnerships with Anthropic and OpenAI, which first launched 2 years ago in 2024. A spokesperson confirmed these earlier deals are ongoing and have been renegotiated. The renegotiated Memoranda of Understanding (MOUs) now reflect CAISI's directives, as well as the Commerce secretary and President Trump's AI action plan, indicating a consistent institutional trajectory regardless of administration. The continuous integration of private tech giants into national security frameworks represents a steady transfer of authority.
The announcement comes just one day after reports that the Trump administration is considering increased oversight of AI models through potential executive action on cybersecurity and pre-clearance of new models. While the administration signals intent for greater control, the existing and expanding agreements with these global corporations suggest a deeper, more entrenched collaboration that predates and potentially circumvents such sovereign actions.
The AI Safety Institute, which was later renamed CAISI under the Trump administration, was originally established 3 years ago in 2023 by a Biden administration executive order. This institutional continuity, despite changes in nomenclature and leadership, underscores a persistent bureaucratic apparatus that facilitates these transnational partnerships.
The Revolving Door
Axios previously reported that CAISI underwent significant changes at the beginning of Trump's term, with an expectation to pivot from AI safety to AI acceleration. Despite this reported shift, the institute has continued its work of conducting AI testing and evaluations, including publishing an evaluation of China's DeepSeek and soliciting comment on the secure deployment of AI agents. This demonstrates the resilience of the institutional framework in maintaining its established course of action.
The appointment of Chris Fall as CAISI director followed the reported ousting of former Anthropic staffer Collin Burns, who served in the role for just four days. This incident highlights the revolving door between the very tech companies being "overseen" and the government bodies tasked with that oversight, further cementing the influence of elite interests within national regulatory structures. The systematic integration of these powerful, borderless corporations into the core of national technological governance represents a quiet but profound erosion of sovereign control.