Strategy and governance for adversarial artificial intelligence

(a) Lead official (1) Designation Not later than March 1, 2027, the Secretary of Defense shall designate a senior official within the Office of the Secretary of Defense to act as the lead for the Department of Defense on matters relating to adversarial artificial intelligence. (2) Coordination The senior official designated pursuant to paragraph (1) shall be responsible for coordinating all Department activities, programs, and investments as they pertain to adversarial artificial intelligence. (3) Notice Not later than 30 days after the date on which the Secretary designates a senior official pursuant to paragraph (1), the Secretary shall submit to the congressional defense committees notice regarding the designation. (b) Strategy and governance structure Not later than January 15, 2028, the Secretary shall, acting through the official designated under subsection (a)(1), establish— (1) a strategy for adversarial artificial intelligence that identifies capability gaps, prioritizes investments, and establishes metrics and milestones for implementation; and (2) a governance structure, chaired by the designated official, that meets not less frequently than quarterly and includes an industry information-sharing mechanism utilizing authorities under section 2224 of title 10, United States Code, and the Defense Industrial Base Cybersecurity program under subpart 236 of title 32, Code of Federal Regulations, as in effect on the day before the date of the enactment of this Act. (c) Briefing Not later than March 1, 2028, and annually thereafter until December 31, 2030, the Secretary shall provide to the congressional defense committees a briefing, at the appropriate level of classification, on administration of this section, including the extent to which the strategy and governance structure established under subsection (b) address the ability of the Department to continuously monitor, in real time, artificial intelligence and machine learning systems in operational use for signs of adversarial manipulation, malfunction, or other unintended behavior. (d) Definitions In this section: (1) The term adversarial artificial intelligence means techniques by which an actor deliberately manipulates, deceives, or exploits artificial intelligence or machine learning systems, including through attacks described in the report National Institute of Standards and Technology Trustworthy and Responsible Artificial Intelligence 100–2e2025 (relating to Adversarial Machine Learning), such as data poisoning, model evasion, model inversion, or supply chain compromise, to cause malfunction or unintended behavior. (2) The term artificial intelligence has the meaning given in section 238(g) of the John S. McCain National Defense Authorization Act for Fiscal Year 2019 (Public Law 115–232; 10 U.S.C. note prec. 4061).