A group of House Democrats is calling on the White House’s Office of Management and Budget (OMB) to mandate Federal agencies adopt the National Institute of Standards and Technology’s (NIST) AI Risk Management Framework (RMF).
In a July 20 letter, Reps. Zoe Lofgren, D-Calif., Ted Lieu, D-Calif., and Haley Stevens, D-Mich., argued that with the rapid development of AI, the Federal government needs to ensure that the technology is developed responsibly and that individuals are protected.
The NIST AI framework – which the agency published at the beginning of the year – details four key functions which the agency considers key to building responsible AI systems: govern, map, measure, and manage. The framework is a roadmap that provides a starting point for organizations deciding how to regulate their use of the technology.
Federal agencies can currently adopt the framework voluntarily.
According to the three House members, requiring Federal agencies to adopt NIST’s AI guidance to analyze and mitigate the risks associated with the technology could significantly affect how the government designs and develops AI systems.
“In its RMF, NIST examines the ways in which organizations can incorporate trustworthiness considerations into the design, development, use, and evaluation of AI systems. The document provides a great starting point for both agencies and vendors to think about AI risk and how their systems can be designed with these risks in mind,” the letter reads.
In addition, the lawmakers acknowledge that Biden Administration has in recent months taken significant steps to hold Federal agencies responsible for addressing bias possibly embedded within AI systems while also promoting innovation. For example, in October 2022, the administration published its AI Bill of Rights blueprint, later followed by the NIST AI framework.
“Through the AI Bill of Rights, the administration highlighted principles that are crucial to ensuring AI is aligned with our democratic values and does not threaten our civil rights,” the letter reads. “The administration has shown a much-needed focus on Federal investments in AI research and development, the risks and opportunities of AI in education, and public assessments of highly-capable generative AI systems.”
Given the administration’s ongoing interest in the rapidly developing technology, lawmakers asked that it also “consider utilizing the NIST AI RMF and subsequent risk management guidance specifically tailored for the Federal government” to ensure agencies and vendors meet baseline standards in mitigating risk.
In addition, lawmakers point out how Federal agencies are already required to follow NIST’s standards and guidelines on cybersecurity. With AI, lawmakers ask that the administration follow a similar approach and ensure Federal agencies and government vendors follow NIST’s standards and guidance as a baseline to mitigate the risks and possible harms of the technology.