Artificial Intelligence in the Draft FY 2025 National Defense Authorization Act

As government leaders examine the possibilities and perils of AI, the topic of regulating the technology often becomes part of the discussion. That is why it is no surprise that Congress is using the FY 2025 national defense bill to both expand AI use cases among defense components as well as encourage assessment of AI risk. The National Defense Authorization Act (NDAA) is often used as an opportunity to support and address emerging technology environments.

Two versions of the FY 2025 NDAA exist, one in the House under H.R. 8070, and the second in the Senate in Executive Summary form. The House passed its version of the NDAA on June 14, 2024, and the Senate completed markup of its version. Next, both versions of the bill will move through the reconciliation process to produce one final bill for passage.  

 AI provisions in each bill include:

House:

• Section 332. Directs the Marine Corps to pilot AI/ML solutions for improving military supply chain readiness, budget efficiency and logistics productivity.
• Section 241. Requires the DOD Chief Digital and AI Officer (CDAO) to establish a secure computing and data storage environment called AIxBio to facilitate the testing of AI models trained on biological data.
• Section 622. Calls for a DOD pilot program using AI/ML to calculate the monthly rates of basic allowances for housing in 15 or more military housing areas.
• Section 1051. Instructs the DOD to use AI/ML to identify methods for defending against biological attacks.
• Section 1511. Orders the DOD to investigate the use of GenAI to protect mobile devices.
• Section 1512. Requires the development of a joint partner sharing network that leverages AI for decision-making.

Senate (note that section numbers are not available in the executive summary):

• Requires a new pilot program to optimize AI-enabled software for the workflow and operations of DOD depots, shipyards, and manufacturing facilities.
• Authorizes increased funding for the development of new manufacturing and production techniques for low-cost UAS and the development of distributed AI/ML fusion for such systems.
• Establishes a working group to develop and coordinate an AI initiative among U.S. allies and partners.
• Creates a pilot program for developing near-term use cases and demonstrations of AI/ML for national security biotechnology applications.
• Requires a review of the DOD’s management of AI-related issues.
• Requires a briefing on opportunities for the creation of lateral entry programs for the fields of AI/ML within the military departments.
• Directs a briefing on the testing infrastructure and criteria currently used or needed to validate AI/ML in military systems.
• Requires a plan to ensure that the budgeting process for programs containing AI/ML includes cost estimates for the acquisition and sustainment of data required to train, maintain or improve the AI.
• Establishes an Artificial Intelligence Human Factors Integration Initiative to improve the human usability of AI systems.
• Encourages DOD to expand the Defense Civilian Training Corps program to explore opportunities in emerging technical and acquisition fields such as advanced manufacturing, AI, and intellectual property rights.

The House and Senate version of the NDAA differ in several aspects when it comes to AI, however, both aim to expand AI adoption in various defense applications, strengthen acquisition and funding for AI, and ensure AI-enabled infrastructure is in place to support the technology. Given these similarities, and despite the above provisions being subject to amendment during the reconciliation process, it is likely we will see a strong AI/ML presence in the final bill. Should that be the case, contractors will see that some of these provisions result in AI opportunities for industry.