The newly enacted National Defense Authorization Act (“NDAA”) contains important provisions regarding the development and deployment of artificial intelligence (“AI”) and machine learning technologies, many of which build upon previous legislation introduced in the 116th Congress. The most substantial federal U.S. legislation on AI to date, these provisions will have significant implications in the national security sector and beyond. The measures in the NDAA will coordinate a national strategy on research, development, and deployment of AI, guiding investment and aligning priorities for its use.

President Trump had vetoed the NDAA after its initial passage in December, but the $740 billion NDAA became law over the objection of President Trump’s veto with a rare New Year’s Day Senate vote, 81-13. The House voted to override President Trump’s veto on December 28, on a 322-87 vote.

This post highlights some of the key AI provisions included in the NDAA.

I.  Establishment of the National Artificial Intelligence Initiative

Building on concepts set forth in prior legislation, including the National Artificial Intelligence Initiative Act of 2020 (S. 1558, H.R. 6216) introduced in the 116th Congress, Title E of the NDAA mandates the establishment of a National Artificial Initiative, for the purpose of:

  1. Ensuring continued United States leadership in artificial intelligence research and development;
  2. Leading the world in the development and use of trustworthy artificial intelligence systems in the public and private sectors;
  3. Preparing the workforce for the integration of artificial intelligence systems across all sectors of the economy and society; and
  4. Coordinating ongoing artificial intelligence research, development, and demonstration activities among the civilian agencies, the Department of Defense and the Intelligence Community.

In support of those goals, the AI Initiative activities will include:

  1. Support for research and development through funding, cooperative agreements, testbeds, and access to data and computing resources;
  2. Support for educational programs, including for K-12 and postsecondary and informal programs to prepare the workforce to create, use, and interact with AI systems;
  3. Support for interdisciplinary research, education, and workforce training for students and researchers in order to, among other things, foster interdisciplinary perspectives that will advance research and development responsibly;
  4. Interagency planning and coordination of Federal AI research, development, demonstration, and standards engagement;
  5. Outreach to diverse stakeholders to ensure public input is taken into account;
  6. Leveraging existing Federal investments to advance AI Initiative objectives;
  7. Support for a network of interdisciplinary AI research institutes; and
  8. Support opportunities for international cooperation with strategic allies on AI research and development, assessment, and resources for trustworthy AI systems.

To implement the AI Initiative, the NDAA mandates the creation of a National Artificial Intelligence Initiative Office under the White House Office of Science and Technology Policy (OSTP) to undertake the AI Initiative activities, as well as an interagency National Artificial Intelligence Advisory Committee to coordinate federal activities pertaining to the AI Initiative. In addition, the Secretary of Commerce, in consultation with other government officials, will establish a National Artificial Intelligence Advisory Committee comprised of members who collectively will provide a broad range of expertise and perspectives. The statute requires the Advisory Committee to establish a subcommittee on AI and law enforcement.

II.  Development of Frameworks through the National Institute of Standards

Building on provisions of several pieces of legislation introduced in the 116th Congress, including the National Artificial Intelligence Initiative Act of 2020 (S. 1558, H.R. 6216) and Advancing Artificial Intelligence Research Act of 2020 (S. 3891), the NDAA directs the National Institute of Standards and Technology (“NIST”) to support the development of relevant standards and best practices pertaining to both artificial intelligence and data sharing. To support these efforts, Congress has appropriated $400 million to NIST through FY 2025.

Specifically, the statute directs NIST to:

  1. Advance collaborative frameworks, standards, guidelines, and techniques for AI;
  2. Support the development of a risk-mitigation framework for deploying AI systems;
  3. Support the development of technical standards and guidelines that promote trustworthy AI systems; and
  4. Support the development of technical standards and guidelines for testing bias in AI training data and applications.

In addition, the legislation also grants the Director of NIST the discretion to:

  1. Support the research and development of best practices and voluntary standards for trustworthy AI systems that may include: (a) standards for privacy and security, including for datasets used to train or test AI systems; (b) advanced computer chips and hardware for AI systems; (c) data management and techniques to increase the usability of data; (e) safety and robustness mechanisms; (f) auditing mechanisms and benchmarks; (g) applications to improve other scientific fields and engineering; (h) model and system documentation; and (i) other matters the Director considers critical for trustworthy AI.
  2. Produce standardized, representative, secure, aggregate, and privacy-protected data sets for AI research, development, and use;
  3. Support institutes established by the National Artificial Intelligence Act of 2020 in order to advance, among other things, trustworthy AI voluntary standards and guidelines;
  4. Support the development of consensus voluntary standards, including international standards; and
  5. Enter into contracts to fulfill NIST goals.

Furthermore, NIST is instructed to (1) develop, in collaboration with public and private organizations, a voluntary risk management framework for trustworthy AI, (2) participate in the development of AI standards and specifications, (3) develop, in collaboration with public and private organizations, guidance to assist with voluntary data sharing among a range of organizations, and (4) develop, in collaboration with public and private sector organizations, best practices for datasets used to train AI, including with respect to documentation.

III. Department of Defense Artificial Intelligence Provisions

The NDAA has several AI-related provisions pertaining to the Department of Defense (“DOD”). Most notably, in relation to the Joint Artificial Intelligence Center (“JAIC”), the new law:

  1. Mandates the creation of a Board of Advisors for the JAIC, composed of experts from outside the DOD, appointed by the Defense Secretary;
  2. Establishes acquisition authority for the Director of the JAIC to procure AI systems and technologies, up to $75 million per year;
  3. Modifies the DOD organizational structure such that the Director of the JAIC will report to the Deputy Secretary of Defense, instead of the DOD Chief Information Officer;
  4. Calls for the acceleration of the development, testing, and fielding of new artificial intelligence capabilities;
  5. Requires an assessment and report on whether AI technology acquired by the DOD is developed in an ethically and responsibly sourced manner, including steps taken or resources required to mitigate any deficiencies; and
  6. Modifies the JAIC biannual report to include an assessment pertaining to the development of AI standards, including an evaluation of the need to incorporate AI standards into DOD strategy and doctrine, development of such standards, and collaboration with other governmental and non-governmental organizations.

Other notable DOD AI provisions include:

  1. Mandating an annual review by the Chiefs of Staff of the need for automated and machine learning or artificial intelligence-assisted electronic warfare capabilities;
  2. Permitting the Defense Secretary to establish a steering committee on emerging technology and national security threats; and
  3. Provision of guidance on use of direct hiring processes for artificial intelligence professionals and other data science and software development personnel.

IV.  Department of Energy AI Research Program

The NDAA authorizes $1.2 billion through FY 2025 for a Department of Energy (“DOE”) artificial intelligence research program, identifying seven key areas for research grants, including the analysis and development of standardized data sets and development of trustworthy AI systems.  To support this program, the Energy Secretary is directed to take certain actions, including making infrastructure, hardware, and software investments and collaborating with many stakeholders.  In carrying out the program, DOE is also directed to support technology transfers of artificial intelligence systems in support of society and United States economic competitiveness.

V.  Other Provisions Expanding Research, Development and Deployment of AI

The NDAA includes several other provisions pertaining to AI. For example, it allocates $4.8 billion to the National Science Foundation, which, among other things, will form a task force, in coordination with OSTP, to investigate the establishment of a National AI Research Resource. These provisions follow those in last session’s National AI Research Resource Task Force Act / National Cloud Computing Task Force Act (H.R.7096, S.3890) , and contemplate that the National Artificial Intelligence Research Resource, if established, among other things, may create a shared computing infrastructure for researchers throughout the United States. Similar to provisions of last session’s National Artificial Intelligence Initiative Act of 2020 (S. 1558, H.R. 6216), the NDAA also authorizes NSF to support the development of a network of inter-disciplinary AI research hubs or institutes to focus on challenges for AI systems such as trustworthiness or that focus on particular economic or social sectors.

Please contact our team for more detailed information about the new laws and their legal implications.

Print:
EmailTweetLikeLinkedIn
Photo of Lee Tiedrich Lee Tiedrich

Lee Tiedrich brings together an undergraduate education in electrical engineering and over twenty years of legal experience to assist clients on a broad range of intellectual property and technology transaction matters. Her work spans several industries, including ehealth, life sciences, consumer products, communications…

Lee Tiedrich brings together an undergraduate education in electrical engineering and over twenty years of legal experience to assist clients on a broad range of intellectual property and technology transaction matters. Her work spans several industries, including ehealth, life sciences, consumer products, communications and media. She counsels both private and public companies, as well as venture capital firms and corporate venture groups in their investments. Ms. Tiedrich has extensive experience negotiating complex intellectual property acquisition, licensing, and development agreements, and regularly counsels clients on strategic issues, such as developing and maintaining intellectual property portfolios and evaluating and addressing intellectual property-related assets and risks.

Kayvan Farchadi

Kayvan Farchadi is a summer associate who attends George Washington University Law School.