WASHINGTON, D.C. – U.S. Senator Peter Welch (D-Vt.) and Senator Ben Ray Luján (D-N.M.) recently introduced the Trustworthy by Design Artificial Intelligence (TBD A.I.) Act, legislation that aims to address realistic, short-term challenges created by unregulated A.I. technologies by establishing best practice guidelines for all phases of A.I. development. The legislation would direct the National Institute of Security and Technology (NIST) to establish a trustworthy-by-design framework to more proactively and systematically manage A.I. risk, create definitions for key trustworthiness evaluation criteria, and identify system components that should be evaluated throughout every stage of A.I. product development. 

The Trustworthy by Design A.I. Act would also help ensure that A.I. models are reliable, safe, and secure, and minimize bias impacting legally protected classes and risks associated with system development. Additionally, the TBD A.I. Act would establish new public reporting requirements for NIST. 

“With great power comes great responsibility—we need to ensure A.I. is used safely. That includes taking action to ensure the government remains a leader in the ethical deployment of A.I. and setting clear guidelines that provide Americans with trustworthy and fair services,” said Senator Welch. “This bill will establish best practices and manage risks of A.I.” 

“There is incredibly promising A.I. innovation happening in New Mexico including new initiatives like the New Mexico A.I. Consortium which I’m confident will foster cutting-edge solutions. These innovations have great potential if we establish clear guidelines on how to keep A.I. secure, resilient, transparent, and fair,” said Senator Luján. “That’s why I am proud to introduce the Trustworthy by Design A.I. Act with Senator Welch to accelerate the development of A.I. guidelines that build consumer trust and pave the road for innovation.” 

The Trustworthy by Design A.I. Act is endorsed by Accountable Tech, Encode, Public Citizen, Public Knowledge, and the Transparency Coalition. 

“The rapid evolution of A.I. has shown that we need independent oversight to ensure responsible A.I. development. Without these guardrails, we are left to deal with the ramifications of these unvetted systems,” said Nicole Gill, co-founder and executive director of Accountable Tech. “The Trustworthy by Design Artificial Intelligence Act encourages a proactive approach to improving this technology for the public interest by centering safety and security in design and development of new A.I. systems. We thank Senator Welch for introducing this legislation, and we urge Congress to act swiftly to increase oversight of developing A.I. technologies.”  

“The TBD A.I. Act represents America’s commitment to leading the world in responsible A.I. innovation. By providing our agencies with the means to confidently deploy A.I. systems through rigorous evaluations for safety, security, and fairness, we’re setting the gold standard for trustworthy artificial intelligence while maintaining our competitive edge,” said Adam Billen, VP of Public Policy, Encode. 

“It’s common sense that the government should ensure that the A.I. tools it is using are trustworthy, unbiased, fair and understandable.  But we can’t rely on A.I. companies to deliver such trustworthy tools, without standards in place. The Trustworthy by Design A.I. Act would convert common sense into policy, strengthening government operations and setting a standard for the private market,” said Robert Weissman, Co-President, Public Citizen. 

“When A.I. developers and deployers call their systems safe, reliable, fair, or responsible it is very hard to trust that they actually are; there are no shared standards or guidelines for how to evaluate A.I. systems or to measure vague claims of trustworthiness. The Trustworthy by Design A.I. Act brings much-needed focus and clarity to the nascent and scattered A.I. evaluation standards development work. It marshals all the essential elements of trustworthiness, and it applies across A.I. system components and the complete lifecycle of A.I. systems. Public Knowledge applauds its focus on robust, independent, cooperative, multi-stakeholder verification and evaluation. This bill forges an essential link in the A.I. accountability chain necessary to ensure A.I. systems are truly trustworthy and accountable,” said Nick Garcia, Senior Policy Counsel, Public Knowledge. 

“The comprehensive guidelines called for in the TBD A.I. Act will provide a solid foundation and much-needed transparency for evaluating A.I. systems, setting the standard for trustworthy A.I. implementation and deployment in Federal agencies,” said Jai Jaisimha, Co-founder, Transparency Coalition. 

Senator Welch has led efforts in the Senate to strengthen consumer protections and safety regulations regarding A.I. In March, Senator Welch joined Senator Luján in introducing the Artificial Intelligence Consumer Opt-In, Notification Standards, and Ethical Norms for Training (AI CONSENT Act), legislation that would require online platforms to obtain consumers’ express informed consent before using their personal data to train A.I. models. Senator Welch also joined Senator Michael Bennet (D-Colo.) to introduce the Digital Platform Commission Act, a bill to create an expert federal agency to provide comprehensive regulation of digital platforms to protect consumers, promote competition, and safeguard the public interest. 

Last month, Senator Welch led the introduction of the Transparency and Responsibility for Artificial Intelligence Networks (TRAIN Act), a new bill to help creators—musicians, artists, writers, and others—access the courts to protect their copyrighted works if and when they are used to train generative A.I. models. In October, Welch introduced the Artificial Intelligence Weapons Accountability and Risk Evaluation (AWARE) Act, which would bolster the U.S. Department of Defense’s ethical technology leadership and transparency regarding A.I.-enabled defense systems. 

Learn more about the Trustworthy by Design A.I. Act.  

Read a section-by-section summary of the bill. 

Read the full text of the bill. 

###