July 29, 2024

Dear Chair Cantwell, Ranking Member Cruz, Chairman Lucas, and Ranking Member Lofgren:

As the Senate Committee on Commerce, Science, and Transportation and the House Committee on Science, Space, and Technology consider legislation pertaining to artificial intelligence, we urge you to prioritize authorizing the U.S. Artificial Intelligence Safety Institute (AISI) within the National Institute of Standards and Technology (NIST).

Establishing the AISI on a statutory basis will ensure that companies of all sizes – as well as other interested parties – continue to have a voice in the development of relevant standards and guidelines. This will accelerate the widespread adoption of AI and further ensure the US continues to lead the world in the development of AI standards.

Currently, businesses of all types recognize the potential of AI, but many have refrained from adoption, in part due to concerns regarding implementation risks. The AISI provides a venue to convene the leading experts across industry and government to contribute to the development of voluntary standards that ultimately assist in derisking adoption of AI technologies.

NIST has previously facilitated innovation and trust in markets with the “NIST Privacy Framework,” a voluntary tool which engenders trust among businesses and customers while facilitating innovative uses of data. As businesses across industries have attested, this tool – developed in collaboration with the private sector and designed to increase transparency and preserve user privacy – has served as a “market differentiator” for adopters and help “support a growing and stable economy.”

The AISI is particularly important for those enterprises not primarily engaged in technological activities and which do not possess the wherewithal to develop bespoke benchmarks and protocols to assess AI systems. NIST, which does not possess regulatory authority and has a long history of successfully engaging the private sector, accomplishes this within the AI space primarily through the AISI Consortium (AISIC).

Launched in February 2024, the AISIC consists of over 200 leading organizations from industry, trade groups, government, civil society and more, working together to share knowledge and “develop science-based and empirically backed guidelines and standards for AI measurement and policy.” Public-private collaboration in standards development also increases the likelihood that companies are able to adopt the standards without being overly burdened. U.S. AI companies have access to unparalleled data, talent, and experience in the challenges to developing and deploying these models. There is no one better than American innovators to speak directly to what guidelines will be necessary, practical, and beneficial.

The AISI has already begun contributing to the development of new guidelines on major topics like red-teaming, risk management, cybersecurity, capability evaluations, and more. Widely disseminating this information will be crucial to derisking AI systems – both from a business and a safety standpoint – ensuring companies in every industry are able to reap their benefits. The Institute’s work has been lauded by its industry Consortium members who have described it as “critical to furthering AI innovation in the United States and globally” and as “a significant step forward in ensuring the safe and ethical development of AI technologies.” Without an organization like AISI, the US risks ceding leadership to other countries (allies and adversaries alike) currently engaged in active development of AI metrics, definitions, standards, and testbeds.

Your committees have the opportunity to bolster the U.S. AI ecosystem by enshrining the AISI in statute. We encourage you to grant appropriate authorization for the AISI so that it can confidently develop the AI safety tools and guidelines which serve as the foundation for widespread adoption of the technology. As the Institute articulates, “safety breeds trust, trust provides confidence in adoption, and adoption accelerates innovation.”

As organizations focused on a wide range of AI policy goals, we agree that the mission of NIST AISI to advance responsible innovation of AI systems is an urgent priority that deserves our full support.

Thank you for your consideration of this request.

Sincerely,

Americans for Responsible Innovation (ARI)

Information Technology Industry Council (ITI)

A.Capital Ventures

AI Policy Institute

Alliance for Trust in AI

Amazon

Association for the Advancement of Artificial Intelligence (AAAI)

BABL AI

BSA | The Software Alliance

Capital Technology University

Carnegie Mellon University

Center for Human-Compatible AI, UC Berkeley

Center for AI Policy

Center for AI Safety Action Fund

CivAI

Clarifai

Cohere

Common Crawl Foundation

Drexel University

Duquesne University, Carl G Grefenstette Center for Ethics

EleutherAI

Encode Justice

FAIR Institute

FAR AI

Federation of American Scientists (FAS)

Free Software Foundation

Future of Life Institute

Human Factors and Ergonomics Society

IBM

Inclusive Abundance Initiative

Institute for Progress

ITIF Center for Data Innovation

Leela AI

Lucid Privacy

Meta

Microsoft

Motorola Solutions

National Retail Federation

Nielsen

OpenAI

Palantir

Palo Alto Networks

Public Knowledge

Salesforce

SandboxAQ

SV Angel

Transformative Futures Institute

TrueLaw

Trustible