Washington, D.C. — U.S. Senator Ted Budd (R-N.C.) sent a letter to the National Cyber Director, Sean Cairncross, and the Director of the Office of Science and Technology Policy, Michael Kratsios, calling for the Center for AI Standards and Innovation (CAISI) to resume publishing its findings and evaluations of frontier Artificial Intelligence (AI) models. In the letter, Senator Budd addresses CAISI’s role in identifying potential vulnerabilities and risks associated with Chinese AI models. Additionally, he notes the need to coordinate and leverage both the private and public sectors to evaluate and monitor AI development to advance American leadership in AI.
In the letter, the Senator wrote:
“To continue to build on the groundbreaking innovation of America’s AI industry, I request that you continue to allow CAISI to perform its important mission in conjunction with the AI missions of other components of the government and, to the extent possible, allow its research to be published. More high-quality research conducted by CAISI’s informed experts deepens America’s AI competitive ecosystem for model and benchmark developers. Leveraging the strength of both America’s public and private sector capacity to evaluate, monitor, and understand frontier AI development will only become more critical as industries in every sector of our economy continue to embrace AI technology.”
Read the full letter text HERE or below:
Dear Director Cairncross and Director Kratsios,
I write regarding reports that the Center for AI Standards and Innovation (CAISI) has been directed to cease the publication of its findings from testing and evaluating frontier AI models. This comes on the heels of President Trump’s June 2, 2026, Executive Order (EO) on “Promoting Advanced Artificial Intelligence Innovation and Security.” The EO notably directs the Secretary of the Treasury, the Director of the National Security Agency (NSA), and the Director of the Cybersecurity and Infrastructure Security Agency (CISA) to “develop and maintain a classified benchmarking process” that assess the cyber capability of advanced AI models.
The EO further tasked those agencies and agency components with determining the threshold at which an AI model should be classified as a “covered frontier model” and empowered the NSA director to decide when the sharing of such assessments with AI developers and researchers is appropriate.
The United States continues to benefit from a healthy and competitive ecosystem of AI researchers and benchmarking organizations. Leading AI labs consistently partner with such groups to evaluate capabilities and probe for potential vulnerabilities. Making these findings open and available benefits future researchers and nurtures the entire research ecosystem. While CAISI is still a relatively new component of the National Institute of Standards and Technology (NIST), it is at the center of the federal government’s efforts to help incubate a robust and rich AI research and evaluation environment. By partnering with the key leading AI labs, NIST’s team of experts have established themselves a trusted resource, for America’s top AI innovators.
As the U.S. continues our fight to maintain American leadership in AI, CAISI has also played a leading role in highlighting, in a public fashion, the relative capability of leading Chinese AI models compared to U.S. offerings. Particularly, CAISI has published research on significant security vulnerabilities found within DeepSeek open-source models so AI researchers, developers, and businesses seeking to incorporate AI can avoid unsafe Chinese platforms.
I agree with the Administration’s efforts to secure critical infrastructure and information systems. The NSA’s world-class cybersecurity expertise and resources are critical in addressing the potential significant national security vulnerabilities and implications of leading models. We can’t ignore the risk of exfiltration of technology by our adversaries or how targeted, malicious prompt injections or data poisoning attempts by state actors could disrupt American technology infrastructure.
CAISI’s mission of facilitating collaborative research fosters and expands the capacity of nongovernmental AI research organizations and AI companies to monitor and understand frontier AI developments. By allowing NSA and CAISI to work together, the Administration can help position AI innovators in the private sector more effectively, especially as AI capabilities improve rapidly.
To continue to build on the groundbreaking innovation of America’s AI industry, I request that you continue to allow CAISI to perform its important mission in conjunction with the AI missions of other components of the government and, to the extent possible, allow its research to be published. More high-quality research conducted by CAISI’s informed experts deepens America’s AI competitive ecosystem for model and benchmark developers. Leveraging the strength of both America’s public and private sector capacity to evaluate, monitor, and understand frontier AI development will only become more critical as industries in every sector of our economy continue to embrace AI technology.
Sincerely,
/X/
###
