Advancing America’s Dominance in AI: The 2021 National Defense Authorization Act’s AI Developments

On January 1, 2021, the 116th Congress enacted the Fiscal Year 2021 National Defense Authorization Act (NDAA), which established, among other things, the National Artificial Intelligence Initiative Act of 2020 (Division E, Title LI, §§ 5101-5106) (hereinafter “the Initiative”), a program to award financial assistance to National Artificial Intelligence Research Institutes (Division E, Title LII, § 5201), and provided acquisition authority to the Joint Artificial Intelligence Center (JAIC). The NDAA’s focus on artificial intelligence (AI) is yet another demonstration of the Federal Government’s priority of and commitment to ensuring American dominance in AI, among other emerging technologies. We have reported on the numerous actions taken by the Federal Government to further that end here, here, here, here, and here. 

The Initiative endeavors to ensure that, for the next ten years, the Federal Government focuses on using “trustworthy” AI systems; preparing the U.S. workforce for the use of AI systems across all sectors of the economy and society; and coordinating AI research, development, and demonstration activities amongst the civilian, defense, and intelligence agencies and communities. To carry out this purpose, the Initiative highlights the need for grants and cooperative agreements for AI research and development, collaboration with industry and other diverse stakeholders, and leveraging existing federal investments to advance the Initiative’s objectives. In support of these activities, the Initiative creates a National Artificial Intelligence Initiative Office (NAIIO). The NAIIO will serve as the point of contact for federal AI activities and aims to regularly work with industry and the public. 

The Initiative also creates an Interagency Committee that will coordinate federal programs and activities and develop a strategic plan with goals, priorities, and metrics to evaluate how federal agencies will carry out the Initiative. The Interagency Committee will expect federal agencies to: (1) prioritize federal leadership and investment in AI research, development, and demonstration; (2) support interdisciplinary AI research, development, demonstration, and education with long-term funding; (3) support research and activities focused on AI ethical, legal, environmental, safety, security, bias, and other societal issues that the use of AI may implicate; (4) ensure the availability of “curated, standardized, secure, representative, aggregate, and privacy-protected data sets for artificial intelligence research and development;” (5) ensure AI research and development has the necessary computing, networking, and data facilities; (6) coordinate and support both federal AI education and work force training activities; and (7) support and coordinate the network of AI research institutes established in Section 5201(b) of the NDAA.

Similarly, the Initiative establishes the National Artificial Intelligence Advisory Committee (NAIAC). The NAIAC will consist of individuals from industry, academic institutions, nonprofit and civil societies, and federal laboratories; these individuals will be appointed to the NAIAC by the Secretary of Commerce. The NAIAC will advise the President and the NAIIO on several items, including providing recommendations on the United States’ competitiveness in AI, progress being made through the Initiative, and regulatory or nonregulatory oversight of AI systems.

In addition, the Initiative calls for: (1) the establishment of a Subcommittee on Artificial Intelligence and Law Enforcement; (2) a study on the impact of AI on the workforce, and (3) the establishment of a task force to address the feasibility of establishing and maintaining a National Artificial Intelligence Research Resource. The Subcommittee on Artificial Intelligence and Law Enforcement will provide advice on: (a) bias in the use of AI systems in law enforcement; (b) law enforcement’s access to data and the security parameters for that data; (c) the ability for the Federal Government and industry to take advantage of AI systems for security or law enforcement purposes; and (d) legal standards to ensure that the use of AI systems is consistent with privacy, civil rights and liberties, and disability laws and rights.

The NDAA also established a program that would allow the Secretaries of Energy and Commerce, and the Director of the National Science Foundation, to award financial assistance to the National Artificial Intelligence Research Institutes program. Part of the National Science Foundation, this program enables longer-term research and U.S. leadership in AI through the creation of AI Research Institutes, which are focused on a particular economic or social sector and either address ethical, societal, safety, and security implications for the use of AI in that sector or focus on cross-cutting challenges for AI systems. There are additional requirements for AI Research Institutes, including, for example, that they must have partnerships across public and private organizations and have the ability to create an innovation ecosystem that can translate the Institute’s research into applications and products. Among other things, the AI Research Institutes may use the financial assistance for developing testbeds for AI systems; managing and making curated, standardized, secure, and privacy protected data sets from the public and private sectors available for purposes of testing and training AI systems; and performing research and providing education activities that involve AI systems to solve social, economic, health, scientific, and national security challenges. Financial assistance may be provided for an initial five-year period, with a potential five-year extension based on a merit-review. The NDAA does not provide specific details on how to seek assistance—only that applications should be submitted to an agency head “at such time, in such manner, and containing such information as the agency head may require.” 

Lastly, the NDAA included acquisition authority for the JAIC. Specifically, the NDAA grants the JAIC Director acquisition authority of up to $75 million for new contracts for each year through FY2025. To support the JAIC Director, the NDAA requires the Department of Defense to provide the JAIC with full-time personnel to execute acquisitions and support program management, cost analysis, and other essential procurement functions.

These are welcomed developments in the world of AI. The NDAA affords industry a number of opportunities to get involved in shaping the Federal Government’s use of AI systems, whether through participating in Committees, influencing the regulatory and/or nonregulatory schemes that may be applied to public and private use of AI systems, or researching and developing AI systems with the support of federal financial assistance.