The US AI Safety Institute (AISI) is a US government body housed within the National Institute of Standards and Technology (NIST) at the Department of Commerce. It was established following President Biden’s October 30, 2023 Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, and its leadership was announced by Commerce Secretary Gina Raimondo on February 7, 2024.
The institute’s mandate is to develop the testing, evaluations, and guidelines needed to assess and reduce the risks of advanced AI systems. NIST framed its mission as creating “the science, practice, and policy of AI safety and trustworthiness” through “science-based, and empirically backed guidelines and standards.” Elizabeth Kelly was named the inaugural director and Elham Tabassi, NIST’s chief AI advisor, was named chief technology officer.
The AISI sat alongside the United Kingdom’s AI Safety Institute as part of an emerging international network of such bodies, and it signed research agreements with leading AI developers to gain early access to models for evaluation. Its work focused on unclassified testing of model capabilities, including those that could bear on national security.
The institute’s standing has shifted with US policy. The 2023 executive order that prompted its creation was rescinded in January 2025, and the office’s role and name were subsequently reorganized within NIST around AI standards and national-security evaluations. The AISI nonetheless marked the point at which the US government built a dedicated technical capacity to evaluate frontier AI.