Center for American Progress

CAP Comment Recommends Key Risk Mitigation Strategies for NIST’s Generative AI Companion Resource
Article

CAP Comment Recommends Key Risk Mitigation Strategies for NIST’s Generative AI Companion Resource

The Center for American Progress submitted comments to the National Institute of Standards and Technology recommending strategies to identify and mitigate harms from generative artificial intelligence.

On February 2, 2024, the Center for American Progress’ Adam Conner, vice president of Technology Policy, and Megan Shahi, director of Technology Policy, submitted a response on behalf of CAP to the National Institute of Standards and Technology’s (NIST) request for information related to NIST’s assignments under Sections 4.1, 4.5, and 11 of the executive order concerning artificial intelligence (AI).

The development and deployment of generative AI is happening at a speed and scale that is likely to exceed previous technological deployments. Unfortunately, this rapid deployment has meant that risk management, trust, and safety features that would traditionally develop over time do not currently exist. On October 30, 2023, President Joe Biden signed the executive order on the safe, secure, and trustworthy development and use of AI, which directed NIST to undertake several tasks, including “developing a companion resource to the AI Risk Management Framework, NIST AI 100-1, for generative AI.”

As CAP has previously stated, voluntary risk management frameworks are not a sufficient substitute for necessary AI regulations and legislation. Scholars have noted the shortcoming of a risk management framing for AI. However, as AI legislation and regulation faces an uphill battle in the United States in the immediate future, voluntary frameworks such as the NIST AI Risk Management Framework can be a first step in helping to identify and potentially mitigate harms from generative AI. As NIST carries out the mission assigned to it by the recent AI executive order, the NIST AI Risk Management Framework generative AI companion, and any updated frameworks, should:

  • Incorporate the White House Blueprint for an AI Bill of Rights.
  • Define and include requirements for the responsibilities and risk management for developers of AI models and the first- and third-party deployers of those AI models.
  • Adopt the categories from the draft U.S. Office of Management and Budget’s AI guidance in which AI use is presumed to affect safety and rights, as well as craft risk mitigation strategies for these categories.
  • Prioritize recommendations to address generative AI’s risks to the integrity of elections and democratic processes given the historic number of elections taking place in 2024.

CAP also submitted a copy of its February 2024 report, “Generative AI Should Be Developed and Deployed Responsibly at Every Level for Everyone,” to NIST to provide further resources for our recommendations for developers of AI models and the first- and third-party deployers of AI models.

CAP’s official comment can be viewed here and is now available in the public record here.

The positions of American Progress, and our policy experts, are independent, and the findings and conclusions presented are those of American Progress alone. A full list of supporters is available here. American Progress would like to acknowledge the many generous supporters who make our work possible.

Authors

Adam Conner

Vice President, Technology Policy

Megan Shahi

Director, Technology Policy

Team

Technology Policy

Our team envisions a better internet for all Americans, advancing ideas that protect consumers, defend their rights, and promote equitable growth.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.